Moontae Lee

Joel Jang, Dongkeun Yoon, Sohee Yang et al. · deepmind, uw

Pretrained Language Models (LMs) memorize a vast amount of knowledge during initial pretraining, including information that may violate the privacy of personal lives and identities. Previous work addressing privacy issues for language models has mostly focused on data preprocessing and differential privacy methods, both requiring re-training the underlying LM. We propose knowledge unlearning as an alternative method to reduce privacy risks for LMs post hoc. We show that simply performing gradient ascent on target token sequences is effective at forgetting them with little to no degradation of general language modeling performances for larger LMs; it sometimes even substantially improves the underlying LM with just a few iterations. We also find that sequential unlearning is better than trying to unlearn all the data at once and that unlearning is highly dependent on which kind of data (domain) is forgotten. By showing comparisons with a previous data preprocessing method and a decoding method known to mitigate privacy risks for LMs, we show that unlearning can give a stronger empirical privacy guarantee in scenarios where the data vulnerable to extraction attacks are known a priori while being much more efficient and robust. We release the code and dataset needed to replicate our results at https://github.com/joeljang/knowledge-unlearning.

Joel Jang, Seungone Kim, Seonghyeon Ye et al. · cmu, uw

Recently, Language Models (LMs) instruction-tuned on multiple tasks, also known as multitask-prompted fine-tuning (MT), have shown the capability to generalize to unseen tasks. Previous work has shown that scaling the number of training tasks is the key component in making stronger MT LMs. In this work, we report an unexpected finding that an expert LM fine-tuned on just a single task can outperform an MT LM trained with 300+ different tasks on 11 different unseen datasets and on 13 datasets of the BIG-bench benchmark by a mean accuracy of 3.20% and 1.29%, respectively. This finding casts doubt on the previously held belief that simply scaling the number of tasks makes stronger MT LMs. Leveraging this finding, we further show that this distributed approach of training a separate expert LM per training task instead of a single MT LM for zero-shot inference possesses many benefits including (1) avoiding negative task transfer that often occurs during instruction tuning, (2) being able to continually learn new tasks without having to re-train on previous tasks to avoid catastrophic forgetting, and (3) showing compositional capabilities when merging individual experts together. The code is available at https://github.com/joeljang/ELM.

Mingqian Zheng, Jiaxin Pei, Lajanugen Logeswaran et al. · stanford

Prompting serves as the major way humans interact with Large Language Models (LLM). Commercial AI systems commonly define the role of the LLM in system prompts. For example, ChatGPT uses ``You are a helpful assistant'' as part of its default system prompt. Despite current practices of adding personas to system prompts, it remains unclear how different personas affect a model's performance on objective tasks. In this study, we present a systematic evaluation of personas in system prompts. We curate a list of 162 roles covering 6 types of interpersonal relationships and 8 domains of expertise. Through extensive analysis of 4 popular families of LLMs and 2,410 factual questions, we demonstrate that adding personas in system prompts does not improve model performance across a range of questions compared to the control setting where no persona is added. Nevertheless, further analysis suggests that the gender, type, and domain of the persona can all influence the resulting prediction accuracies. We further experimented with a list of persona search strategies and found that, while aggregating results from the best persona for each question significantly improves prediction accuracy, automatically identifying the best persona is challenging, with predictions often performing no better than random selection. Overall, our findings suggest that while adding a persona may lead to performance gains in certain settings, the effect of each persona can be largely random. Code and data are available at https://github.com/Jiaxin-Pei/Prompting-with-Social-Roles.

Jinwoo Kim, Tien Dat Nguyen, Seonwoo Min et al.

We show that standard Transformers without graph-specific modifications can lead to promising results in graph learning both in theory and practice. Given a graph, we simply treat all nodes and edges as independent tokens, augment them with token embeddings, and feed them to a Transformer. With an appropriate choice of token embeddings, we prove that this approach is theoretically at least as expressive as an invariant graph network (2-IGN) composed of equivariant linear layers, which is already more expressive than all message-passing Graph Neural Networks (GNN). When trained on a large-scale graph dataset (PCQM4Mv2), our method coined Tokenized Graph Transformer (TokenGT) achieves significantly better results compared to GNN baselines and competitive results compared to Transformer variants with sophisticated graph-specific inductive bias. Our implementation is available at https://github.com/jw9730/tokengt.

Sungjun Cho, Seonwoo Min, Jinwoo Kim et al.

To overcome the quadratic cost of self-attention, recent works have proposed various sparse attention modules, most of which fall under one of two groups: 1) sparse attention under a hand-crafted patterns and 2) full attention followed by a sparse variant of softmax such as $α$-entmax. Unfortunately, the first group lacks adaptability to data while the second still requires quadratic cost in training. In this work, we propose SBM-Transformer, a model that resolves both problems by endowing each attention head with a mixed-membership Stochastic Block Model (SBM). Then, each attention head data-adaptively samples a bipartite graph, the adjacency of which is used as an attention mask for each input. During backpropagation, a straight-through estimator is used to flow gradients beyond the discrete sampling step and adjust the probabilities of sampled edges based on the predictive loss. The forward and backward cost are thus linear to the number of edges, which each attention head can also choose flexibly based on the input. By assessing the distribution of graphs, we theoretically show that SBM-Transformer is a universal approximator for arbitrary sequence-to-sequence functions in expectation. Empirical evaluations under the LRA and GLUE benchmarks demonstrate that our model outperforms previous efficient variants as well as the original Transformer with full attention. Our implementation can be found in https://github.com/sc782/SBM-Transformer .

Muhammad Khalifa, Lajanugen Logeswaran, Moontae Lee et al.

We study few-shot reranking for multi-hop QA with open-domain questions. To alleviate the need for a large number of labeled question-document pairs for retriever training, we propose PromptRank, which relies on large language models prompting for multi-hop path reranking. PromptRank first constructs an instruction-based prompt that includes a candidate document path and then computes the relevance score between a given question and the path based on the conditional likelihood of the question given the path prompt according to a language model. PromptRank yields strong retrieval performance on HotpotQA with only 128 training examples compared to state-of-the-art methods trained on thousands of examples -- 73.6 recall@10 by PromptRank vs. 77.8 by PathRetriever and 77.5 by multi-hop dense retrieval. Code available at https://github.com/mukhal/PromptRank

Sungmin Cha, Sungjun Cho, Dasol Hwang et al.

Since the recent advent of regulations for data protection (e.g., the General Data Protection Regulation), there has been increasing demand in deleting information learned from sensitive data in pre-trained models without retraining from scratch. The inherent vulnerability of neural networks towards adversarial attacks and unfairness also calls for a robust method to remove or correct information in an instance-wise fashion, while retaining the predictive performance across remaining data. To this end, we consider instance-wise unlearning, of which the goal is to delete information on a set of instances from a pre-trained model, by either misclassifying each instance away from its original prediction or relabeling the instance to a different label. We also propose two methods that reduce forgetting on the remaining data: 1) utilizing adversarial examples to overcome forgetting at the representation-level and 2) leveraging weight importance metrics to pinpoint network parameters guilty of propagating unwanted information. Both methods only require the pre-trained model and data instances to forget, allowing painless application to real-life settings where the entire training set is unavailable. Through extensive experimentation on various image classification benchmarks, we show that our approach effectively preserves knowledge of remaining data while unlearning given instances in both single-task and continual unlearning scenarios.

Soyoung An, Kyunghoon Bae, Eunbi Choi et al.

We introduce EXAONE 3.0 instruction-tuned language model, the first open model in the family of Large Language Models (LLMs) developed by LG AI Research. Among different model sizes, we publicly release the 7.8B instruction-tuned model to promote open research and innovations. Through extensive evaluations across a wide range of public and in-house benchmarks, EXAONE 3.0 demonstrates highly competitive real-world performance with instruction-following capability against other state-of-the-art open models of similar size. Our comparative analysis shows that EXAONE 3.0 excels particularly in Korean, while achieving compelling performance across general tasks and complex reasoning. With its strong real-world effectiveness and bilingual proficiency, we hope that EXAONE keeps contributing to advancements in Expert AI. Our EXAONE 3.0 instruction-tuned model is available at https://huggingface.co/LGAI-EXAONE/EXAONE-3.0-7.8B-Instruct.

Muhammad Khalifa, Lajanugen Logeswaran, Moontae Lee et al.

In-context learning (ICL) operates by showing language models (LMs) examples of input-output pairs for a given task, i.e., demonstrations. The standard approach for ICL is to prompt the LM with concatenated demonstrations followed by the test input. This approach suffers from some issues. First, concatenation offers almost no control over the contribution of each demo to the model prediction. This can be sub-optimal when some demonstrations are irrelevant to the test example. Second, due to the input length limit of some transformer models, it might be infeasible to fit many examples into the context, especially when dealing with long-input tasks. In this work, we explore Demonstration Ensembling (DENSE) as an alternative to simple concatenation. DENSE predicts outputs using subsets (i.e., buckets) of the demonstrations and then combines the output probabilities resulting from each subset to produce the final prediction. We study different ensembling methods using GPT-j and experiment on 12 language tasks. Our experiments show weighted max ensembling to outperform vanilla concatenation by as large as 2.4 average points. Code available at https://github.com/mukhal/icl-ensembling.

Sungmin Cha, Sungjun Cho, Dasol Hwang et al.

Large Language Models (LLMs) have demonstrated strong reasoning and memorization capabilities via pretraining on massive textual corpora. However, this poses risk of privacy and copyright violations, highlighting the need for efficient machine unlearning methods that remove sensitive data without retraining from scratch. While Gradient Ascent (GA) is commonly used to unlearn by reducing the likelihood of generating unwanted content, it leads to unstable optimization and catastrophic forgetting of retrained knowledge. We find that combining GA with low-rank adaptation results in poor trade-offs between computational cost and generative performance. To address these challenges, we propose Low-rank Knowledge Unlearning (LoKU), a novel framework that enables robust and efficient unlearning for LLMs. First, we introduce Inverted Hinge Loss, which suppresses unwanted tokens while maintaining fluency by boosting the probability of the next most likely token. Second, we develop a data-adaptive initialization for LoRA adapters via low-rank approximation weighted with relative Fisher information, thereby focusing updates on parameters critical for removing targeted knowledge. Experiments on the Training Data Extraction Challenge dataset using GPT-Neo models as well as on the TOFU benchmark with Phi-1.5B and Llama2-7B models demonstrate that our approach effectively removes sensitive information while maintaining reasoning and generative capabilities with minimal impact. Our implementation can be found in https://github.com/csm9493/efficient-llm-unlearning.

Byoungjip Kim, Sungik Choi, Dasol Hwang et al.

Despite surprising performance on zero-shot transfer, pre-training a large-scale multimodal model is often prohibitive as it requires a huge amount of data and computing resources. In this paper, we propose a method (BeamCLIP) that can effectively transfer the representations of a large pre-trained multimodal model (CLIP-ViT) into a small target model (e.g., ResNet-18). For unsupervised transfer, we introduce cross-modal similarity matching (CSM) that enables a student model to learn the representations of a teacher model by matching the relative similarity distribution across text prompt embeddings. To better encode the text prompts, we design context-based prompt augmentation (CPA) that can alleviate the lexical ambiguity of input text prompts. Our experiments show that unsupervised representation transfer of a pre-trained vision-language model enables a small ResNet-18 to achieve a better ImageNet-1K top-1 linear probe accuracy (66.2%) than vision-only self-supervised learning (SSL) methods (e.g., SimCLR: 51.8%, SwAV: 63.7%), while closing the gap with supervised learning (69.8%).

Bangzhao Shu, Lechen Zhang, Minje Choi et al.

The versatility of Large Language Models (LLMs) on natural language understanding tasks has made them popular for research in social sciences. To properly understand the properties and innate personas of LLMs, researchers have performed studies that involve using prompts in the form of questions that ask LLMs about particular opinions. In this study, we take a cautionary step back and examine whether the current format of prompting LLMs elicits responses in a consistent and robust manner. We first construct a dataset that contains 693 questions encompassing 39 different instruments of persona measurement on 115 persona axes. Additionally, we design a set of prompts containing minor variations and examine LLMs' capabilities to generate answers, as well as prompt variations to examine their consistency with respect to content-level variations such as switching the order of response options or negating the statement. Our experiments on 17 different LLMs reveal that even simple perturbations significantly downgrade a model's question-answering ability, and that most LLMs have low negation consistency. Our results suggest that the currently widespread practice of prompting is insufficient to accurately and reliably capture model perceptions, and we therefore discuss potential alternatives to improve these issues.

Sungjun Cho, Dae-Woong Jeong, Sung Moon Ko et al.

Pretraining molecular representations from large unlabeled data is essential for molecular property prediction due to the high cost of obtaining ground-truth labels. While there exist various 2D graph-based molecular pretraining approaches, these methods struggle to show statistically significant gains in predictive performance. Recent work have thus instead proposed 3D conformer-based pretraining under the task of denoising, which led to promising results. During downstream finetuning, however, models trained with 3D conformers require accurate atom-coordinates of previously unseen molecules, which are computationally expensive to acquire at scale. In light of this limitation, we propose D&D, a self-supervised molecular representation learning framework that pretrains a 2D graph encoder by distilling representations from a 3D denoiser. With denoising followed by cross-modal knowledge distillation, our approach enjoys use of knowledge obtained from denoising as well as painless application to downstream tasks with no access to accurate conformers. Experiments on real-world molecular property prediction datasets show that the graph encoder trained via D&D can infer 3D information based on the 2D graph and shows superior performance and label-efficiency against other baselines.

Yunxiang Zhang, Muhammad Khalifa, Lajanugen Logeswaran et al.

Open-domain question answering (QA) systems are often built with retrieval modules. However, retrieving passages from a given source is known to suffer from insufficient knowledge coverage. Alternatively, prompting large language models (LLMs) to generate contextual passages based on their parametric knowledge has been shown to improve QA performance. Yet, LLMs tend to "hallucinate" content that conflicts with the retrieved knowledge. Based on the intuition that answers supported by both sources are more likely to be correct, we propose COMBO, a Compatibility-Oriented knowledge Merging for Better Open-domain QA framework, to effectively leverage the two sources of information. Concretely, we match LLM-generated passages with retrieved counterparts into compatible pairs, based on discriminators trained with silver compatibility labels. Then a Fusion-in-Decoder-based reader model handles passage pairs to arrive at the final answer. Experiments show that COMBO outperforms competitive baselines on three out of four tested open-domain QA benchmarks. Further analysis reveals that our proposed framework demonstrates greater efficacy in scenarios with a higher degree of knowledge conflicts.

Lajanugen Logeswaran, Yao Fu, Moontae Lee et al.

Pre-trained large language models have shown successful progress in many language understanding benchmarks. This work explores the capability of these models to predict actionable plans in real-world environments. Given a text instruction, we show that language priors encoded in pre-trained language models allow us to infer fine-grained subgoal sequences. In contrast to recent methods which make strong assumptions about subgoal supervision, our experiments show that language models can infer detailed subgoal sequences from few training sequences without any fine-tuning. We further propose a simple strategy to re-rank language model predictions based on interaction and feedback from the environment. Combined with pre-trained navigation and visual reasoning components, our approach demonstrates competitive performance on subgoal prediction and task completion in the ALFRED benchmark compared to prior methods that assume more subgoal supervision.

Kyungjae Lee, Sang-eun Han, Seung-won Hwang et al.

This paper studies the problem of open-domain question answering, with the aim of answering a diverse range of questions leveraging knowledge resources. Two types of sources, QA-pair and document corpora, have been actively leveraged with the following complementary strength. The former is highly precise when the paraphrase of given question $q$ was seen and answered during training, often posed as a retrieval problem, while the latter generalizes better for unseen questions. A natural follow-up is thus leveraging both models, while a naive pipelining or integration approaches have failed to bring additional gains over either model alone. Our distinction is interpreting the problem as calibration, which estimates the confidence of predicted answers as an indicator to decide when to use a document or QA-pair corpus. The effectiveness of our method was validated on widely adopted benchmarks such as Natural Questions and TriviaQA.

Seungyeon Rhyu, Kichang Yang, Sungjun Cho et al.

Music generation introduces challenging complexities to large language models. Symbolic structures of music often include vertical harmonization as well as horizontal counterpoint, urging various adaptations and enhancements for large-scale Transformers. However, existing works share three major drawbacks: 1) their tokenization requires domain-specific annotations, such as bars and beats, that are typically missing in raw MIDI data; 2) the pure impact of enhancing token embedding methods is hardly examined without domain-specific annotations; and 3) existing works to overcome the aforementioned drawbacks, such as MuseNet, lack reproducibility. To tackle such limitations, we develop a MIDI-based music generation framework inspired by MuseNet, empirically studying two structural embeddings that do not rely on domain-specific annotations. We provide various metrics and insights that can guide suitable encoding to deploy. We also verify that multiple embedding configurations can selectively boost certain musical aspects. By providing open-source implementations via HuggingFace, our findings shed light on leveraging large language models toward practical and reproducible music generation.

Seungone Kim, Juyoung Suk, Shayne Longpre et al. · cmu

Proprietary LMs such as GPT-4 are often employed to assess the quality of responses from various LMs. However, concerns including transparency, controllability, and affordability strongly motivate the development of open-source LMs specialized in evaluations. On the other hand, existing open evaluator LMs exhibit critical shortcomings: 1) they issue scores that significantly diverge from those assigned by humans, and 2) they lack the flexibility to perform both direct assessment and pairwise ranking, the two most prevalent forms of assessment. Additionally, they do not possess the ability to evaluate based on custom evaluation criteria, focusing instead on general attributes like helpfulness and harmlessness. To address these issues, we introduce Prometheus 2, a more powerful evaluator LM than its predecessor that closely mirrors human and GPT-4 judgements. Moreover, it is capable of processing both direct assessment and pair-wise ranking formats grouped with a user-defined evaluation criteria. On four direct assessment benchmarks and four pairwise ranking benchmarks, Prometheus 2 scores the highest correlation and agreement with humans and proprietary LM judges among all tested open evaluator LMs. Our models, code, and data are all publicly available at https://github.com/prometheus-eval/prometheus-eval.

Yunseok Jang, Sungryull Sohn, Lajanugen Logeswaran et al.

Real-world tasks consist of multiple inter-dependent subtasks (e.g., a dirty pan needs to be washed before it can be used for cooking). In this work, we aim to model the causal dependencies between such subtasks from instructional videos describing the task. This is a challenging problem since complete information about the world is often inaccessible from videos, which demands robust learning mechanisms to understand the causal structure of events. We present Multimodal Subtask Graph Generation (MSG2), an approach that constructs a Subtask Graph defining the dependency between a task's subtasks relevant to a task from noisy web videos. Graphs generated by our multimodal approach are closer to human-annotated graphs compared to prior approaches. MSG2 further performs the downstream task of next subtask prediction 85% and 30% more accurately than recent video transformer models in the ProceL and CrossTask datasets, respectively.

Sungjun Cho, Seunghyuk Cho, Sungwoo Park et al.

Real-world graphs naturally exhibit hierarchical or cyclical structures that are unfit for the typical Euclidean space. While there exist graph neural networks that leverage hyperbolic or spherical spaces to learn representations that embed such structures more accurately, these methods are confined under the message-passing paradigm, making the models vulnerable against side-effects such as oversmoothing and oversquashing. More recent work have proposed global attention-based graph Transformers that can easily model long-range interactions, but their extensions towards non-Euclidean geometry are yet unexplored. To bridge this gap, we propose Fully Product-Stereographic Transformer, a generalization of Transformers towards operating entirely on the product of constant curvature spaces. When combined with tokenized graph Transformers, our model can learn the curvature appropriate for the input graph in an end-to-end fashion, without the need of additional tuning on different curvature initializations. We also provide a kernelized approach to non-Euclidean attention, which enables our model to run in time and memory cost linear to the number of nodes and edges while respecting the underlying geometry. Experiments on graph reconstruction and node classification demonstrate the benefits of generalizing Transformers to the non-Euclidean domain.

Lajanugen Logeswaran, Sungryull Sohn, Yunseok Jang et al.

This work explores the problem of generating task graphs of real-world activities. Different from prior formulations, we consider a setting where text transcripts of instructional videos performing a real-world activity (e.g., making coffee) are provided and the goal is to identify the key steps relevant to the task as well as the dependency relationship between these key steps. We propose a novel task graph generation approach that combines the reasoning capabilities of instruction-tuned language models along with clustering and ranking components to generate accurate task graphs in a completely unsupervised manner. We show that the proposed approach generates more accurate task graphs compared to a supervised learning approach on tasks from the ProceL and CrossTask datasets.

Sung Moon Ko, Sungjun Cho, Dae-Woong Jeong et al.

Graph pooling is a crucial operation for encoding hierarchical structures within graphs. Most existing graph pooling approaches formulate the problem as a node clustering task which effectively captures the graph topology. Conventional methods ask users to specify an appropriate number of clusters as a hyperparameter, then assume that all input graphs share the same number of clusters. In inductive settings where the number of clusters can vary, however, the model should be able to represent this variation in its pooling layers in order to learn suitable clusters. Thus we propose GMPool, a novel differentiable graph pooling architecture that automatically determines the appropriate number of clusters based on the input data. The main intuition involves a grouping matrix defined as a quadratic form of the pooling operator, which induces use of binary classification probabilities of pairwise combinations of nodes. GMPool obtains the pooling operator by first computing the grouping matrix, then decomposing it. Extensive evaluations on molecular property prediction tasks demonstrate that our method outperforms conventional methods.

Sungmin Cha, Jihwan Kwak, Dongsub Shim et al.

Class incremental learning (CIL) algorithms aim to continually learn new object classes from incrementally arriving data while not forgetting past learned classes. The common evaluation protocol for CIL algorithms is to measure the average test accuracy across all classes learned so far -- however, we argue that solely focusing on maximizing the test accuracy may not necessarily lead to developing a CIL algorithm that also continually learns and updates the representations, which may be transferred to the downstream tasks. To that end, we experimentally analyze neural network models trained by CIL algorithms using various evaluation protocols in representation learning and propose new analysis methods. Our experiments show that most state-of-the-art algorithms prioritize high stability and do not significantly change the learned representation, and sometimes even learn a representation of lower quality than a naive baseline. However, we observe that these algorithms can still achieve high test accuracy because they enable a model to learn a classifier that closely resembles an estimated linear classifier trained for linear probing. Furthermore, the base model learned in the first task, which involves single-task learning, exhibits varying levels of representation quality across different algorithms, and this variance impacts the final performance of CIL algorithms. Therefore, we suggest that the representation-level evaluation should be considered as an additional recipe for more diverse evaluation for CIL algorithms.

Zheyuan Zhang, Shane Storks, Fengyuan Hu et al.

Pre-trained language models (PLMs) have shown impressive performance in various language tasks. However, they are prone to spurious correlations, and often generate illusory information. In real-world applications, PLMs should justify decisions with formalized, coherent reasoning chains, but this challenge remains under-explored. Cognitive psychology theorizes that humans are capable of utilizing fast and intuitive heuristic thinking to make decisions based on past experience, then rationalizing the decisions through slower and deliberative analytic reasoning. We incorporate these interlinked dual processes in fine-tuning and in-context learning with PLMs, applying them to two language understanding tasks that require coherent physical commonsense reasoning. We show that our proposed Heuristic-Analytic Reasoning (HAR) strategies drastically improve the coherence of rationalizations for model decisions, yielding state-of-the-art results on Tiered Reasoning for Intuitive Physics (TRIP). We also find that this improved coherence is a direct result of more faithful attention to relevant language context in each step of reasoning. Our findings suggest that human-like reasoning strategies can effectively improve the coherence and reliability of PLM reasoning.

Lajanugen Logeswaran, Sungryull Sohn, Yiwei Lyu et al.

One of the fundamental skills required for an agent acting in an environment to complete tasks is the ability to understand what actions are plausible at any given point. This work explores a novel use of code representations to reason about action preconditions for sequential decision making tasks. Code representations offer the flexibility to model procedural activities and associated constraints as well as the ability to execute and verify constraint satisfaction. Leveraging code representations, we extract action preconditions from demonstration trajectories in a zero-shot manner using pre-trained code models. Given these extracted preconditions, we propose a precondition-aware action sampling strategy that ensures actions predicted by a policy are consistent with preconditions. We demonstrate that the proposed approach enhances the performance of few-shot policy learning approaches across task-oriented dialog and embodied textworld benchmarks.

Muhammad Khalifa, Lajanugen Logeswaran, Jaekyeom Kim et al.

Large language models (LLMs) are increasingly used as judges to evaluate agent performance, particularly in non-verifiable settings where judgments rely on agent trajectories including chain-of-thought (CoT) reasoning. This paradigm implicitly assumes that the agent's CoT faithfully reflects both its internal reasoning and the underlying environment state. We show this assumption is brittle: LLM judges are highly susceptible to manipulation of agent reasoning traces. By systematically rewriting agent CoTs while holding actions and observations fixed, we demonstrate that manipulated reasoning alone can inflate false positive rates of state-of-the-art VLM judges by up to 90% across 800 trajectories spanning diverse web tasks. We study manipulation strategies spanning style-based approaches that alter only the presentation of reasoning and content-based approaches that fabricate signals of task progress, and find that content-based manipulations are consistently more effective. We evaluate prompting-based techniques and scaling judge-time compute, which reduce but do not fully eliminate susceptibility to manipulation. Our findings reveal a fundamental vulnerability in LLM-based evaluation and highlight the need for judging mechanisms that verify reasoning claims against observable evidence.

Muhammad Khalifa, Rishabh Agarwal, Lajanugen Logeswaran et al.

Step-by-step verifiers -- also known as process reward models (PRMs) -- are a key ingredient for test-time scaling. PRMs require step-level supervision, making them expensive to train. This work aims to build data-efficient PRMs as verbalized step-wise reward models that verify every step in the solution by generating a verification chain-of-thought (CoT). We propose ThinkPRM, a long CoT verifier fine-tuned on orders of magnitude fewer process labels than those required by discriminative PRMs. Our approach capitalizes on the inherent reasoning abilities of long CoT models, and outperforms LLM-as-a-Judge and discriminative verifiers -- using only 1% of the process labels in PRM800K -- across several challenging benchmarks. Specifically, ThinkPRM beats the baselines on ProcessBench, MATH-500, and AIME '24 under best-of-N selection and reward-guided search. In an out-of-domain evaluation on a subset of GPQA-Diamond and LiveCodeBench, our PRM surpasses discriminative verifiers trained on the full PRM800K by 8% and 4.5%, respectively. Lastly, under the same token budget, ThinkPRM scales up verification compute more effectively compared to LLM-as-a-Judge, outperforming it by 7.2% on a subset of ProcessBench. Our work highlights the value of generative, long CoT PRMs that can scale test-time compute for verification while requiring minimal supervision for training. Our code, data, and models are released at https://github.com/mukhal/thinkprm.

Lechen Zhang, Yusheng Zhou, Tolga Ergen et al.

System prompts provide a lightweight yet powerful mechanism for conditioning large language models (LLMs) at inference time. While prior work has focused on English-only settings, real-world deployments benefit from having a single prompt to operate reliably across languages. This paper presents a comprehensive study of how different system prompts steer models toward accurate and robust cross-lingual behavior. We propose a unified four-dimensional evaluation framework to assess system prompts in multilingual environments. Through large-scale experiments on five languages, three LLMs, and three benchmarks, we uncover that certain prompt components, such as CoT, emotion, and scenario, correlate with robust multilingual behavior. We develop a prompt optimization framework for multilingual settings and show it can automatically discover prompts that improve all metrics by 5-10%. Finally, we analyze over 10 million reasoning units and find that more performant system prompts induce more structured and consistent reasoning patterns, while reducing unnecessary language-switching. Together, we highlight system prompt optimization as a scalable path to accurate and robust multilingual LLM behavior.

Jiyeon Kim, Sungik Choi, Yongrae Jo et al.

Diffusion-based language models (dLLMs) have emerged as a promising alternative to autoregressive language models, offering the potential for parallel token generation and bidirectional context modeling. However, harnessing this flexibility for fully non-autoregressive decoding remains an open question, particularly for reasoning and planning tasks. In this work, we investigate non-autoregressive decoding in dLLMs by systematically analyzing its inference dynamics along the temporal axis. Specifically, we uncover an inherent failure mode in confidence-based non-autoregressive generation stemming from a strong proximity bias-the tendency for the denoising order to concentrate on spatially adjacent tokens. This local dependency leads to spatial error propagation, rendering the entire trajectory critically contingent on the initial unmasking position. Leveraging this insight, we present a minimal-intervention approach that guides early token selection, employing a lightweight planner and end-of-sequence temperature annealing. We thoroughly evaluate our method on various reasoning and planning tasks and observe substantial overall improvement over existing heuristic baselines without significant computational overhead.

Jaehoon Lee, Hankook Lee, Sungik Choi et al.

When solving forecasting problems including multiple time-series features, existing approaches often fall into two extreme categories, depending on whether to utilize inter-feature information: univariate and complete-multivariate models. Unlike univariate cases which ignore the information, complete-multivariate models compute relationships among a complete set of features. However, despite the potential advantage of leveraging the additional information, complete-multivariate models sometimes underperform univariate ones. Therefore, our research aims to explore a middle ground between these two by introducing what we term Partial-Multivariate models where a neural network captures only partial relationships, that is, dependencies within subsets of all features. To this end, we propose PMformer, a Transformer-based partial-multivariate model, with its training algorithm. We demonstrate that PMformer outperforms various univariate and complete-multivariate models, providing a theoretical rationale and empirical analysis for its superiority. Additionally, by proposing an inference technique for PMformer, the forecasting accuracy is further enhanced. Finally, we highlight other advantages of PMformer: efficiency and robustness under missing features.

Saelyne Yang, Sunghyun Park, Yunseok Jang et al.

Instructional videos provide detailed how-to guides for various tasks, with viewers often posing questions regarding the content. Addressing these questions is vital for comprehending the content, yet receiving immediate answers is difficult. While numerous computational models have been developed for Video Question Answering (Video QA) tasks, they are primarily trained on questions generated based on video content, aiming to produce answers from within the content. However, in real-world situations, users may pose questions that go beyond the video's informational boundaries, highlighting the necessity to determine if a video can provide the answer. Discerning whether a question can be answered by video content is challenging due to the multi-modal nature of videos, where visual and verbal information are intertwined. To bridge this gap, we present the YTCommentQA dataset, which contains naturally-generated questions from YouTube, categorized by their answerability and required modality to answer -- visual, script, or both. Experiments with answerability classification tasks demonstrate the complexity of YTCommentQA and emphasize the need to comprehend the combined role of visual and script information in video reasoning. The dataset is available at https://github.com/lgresearch/YTCommentQA.

Hojae Han, Jaejin Kim, Seung-won Hwang et al.

This work addresses test output prediction, a key challenge in test case generation. To improve the reliability of predicted outputs by LLMs, prior approaches generate code first to ground predictions. One grounding strategy is direct execution of generated code, but even minor errors can cause failures. To address this, we introduce LLM-based pseudocode execution, which grounds prediction on more error-resilient pseudocode and simulates execution via LLM reasoning. We further propose DuET, a dual-execution framework that combines both approaches by functional majority voting. Our analysis shows the two approaches are complementary in overcoming the limitations of direct execution suffering from code errors, and pseudocode reasoning from hallucination. On LiveCodeBench, DuET achieves the state-of-the-art performance, improving Pass@1 by 13.6 pp.

Jongho Kim, Jaeyoung Kim, Seung-won Hwang et al.

We study leveraging adaptive retrieval to ensure sufficient "bridge" documents are retrieved for reasoning-intensive retrieval. Bridge documents are those that contribute to the reasoning process yet are not directly relevant to the initial query. While existing reasoning-based reranker pipelines attempt to surface these documents in ranking, they suffer from bounded recall. Naive solution with adaptive retrieval into these pipelines often leads to planning error propagation. To address this, we propose REPAIR, a framework that bridges this gap by repurposing reasoning plans as dense feedback signals for adaptive retrieval. Our key distinction is enabling mid-course correction during reranking through selective adaptive retrieval, retrieving documents that support the pivotal plan. Experimental results on reasoning-intensive retrieval and complex QA tasks demonstrate that our method outperforms existing baselines by 5.6%pt.

Dahyun Lee, Yongrae Jo, Haeju Park et al.

Retrieval in Retrieval-Augmented Generation(RAG) must ensure that retrieved passages are not only individually relevant but also collectively form a comprehensive set. Existing approaches primarily rerank top-k passages based on their individual relevance, often failing to meet the information needs of complex queries in multi-hop question answering. In this work, we propose a set-wise passage selection approach and introduce SETR, which explicitly identifies the information requirements of a query through Chain-of-Thought reasoning and selects an optimal set of passages that collectively satisfy those requirements. Experiments on multi-hop RAG benchmarks show that SETR outperforms both proprietary LLM-based rerankers and open-source baselines in terms of answer correctness and retrieval quality, providing an effective and efficient alternative to traditional rerankers in RAG systems. The code is available at https://github.com/LGAI-Research/SetR

Takyoung Kim, Kyungjae Lee, Young Rok Jang et al.

Interactions with large language models (LLMs) often yield long and detailed responses, leveraging both parametric knowledge and retrieval-augmented generation (RAG). While these responses can provide rich insights, they often include redundant or less engaging content not aligned with user interests. This issue becomes apparent when users specify particular subtopics to include or exclude -- termed coverage-conditioned ($C^2$) queries -- as LLMs often struggle to provide tailored responses. To address this challenge, we investigate the role of query outlines, sequences of subqueries designed to guide LLMs in generating responses that meet specific user requirements. To systematically create and evaluate these outlines, we introduce QTree, a dataset of 10K hierarchical sets of information-seeking subqueries that define structured boundaries for outline creation and evaluation in $C^2$ scenarios. Additionally, we develop QPlanner, a 7B language model trained to generate customized outlines within boundaries of QTree. We evaluate the effectiveness of the generated outlines through automatic and human judgements, focusing on their impact within retrieval-augmented generation (RAG) systems. Experimental results demonstrate that QPlanner, especially when trained with alignment techniques like DPO, generates higher-quality outlines that better fulfill diverse user needs.

Yunxiang Zhang, Muhammad Khalifa, Shitanshu Bhushan et al.

We introduce MLRC-Bench, a benchmark designed to quantify how effectively language agents can tackle challenging Machine Learning (ML) Research Competitions, with a focus on open research problems that demand novel methodologies. Unlike prior work, e.g., AI Scientist, which evaluates the end-to-end agentic pipeline by using LLM-as-a-judge, MLRC-Bench measures the key steps of proposing and implementing novel research methods and evaluates them with rigorous protocol and objective metrics. Our curated suite of 7 competition tasks reveals significant challenges for LLM agents. Even the best-performing tested agent (gemini-exp-1206 under MLAB) closes only 9.3% of the gap between baseline and top human participant scores. Furthermore, our analysis reveals a misalignment between the LLM-judged innovation and actual performance on cutting-edge ML research problems. MLRC-Bench is a dynamic benchmark, designed to grow with new ML competitions and encourage rigorous, objective evaluations of AI research capabilities. Our leaderboard and code are available at: https://huggingface.co/spaces/launch/MLRC_Bench

Seongyun Lee, Yongrae Jo, Minju Seo et al.

Recent advances in reasoning models and agentic AI systems have led to an increased reliance on diverse external information. However, this shift introduces input contexts that are inherently noisy, a reality that current sanitized benchmarks fail to capture. We introduce NoisyBench, a comprehensive benchmark that systematically evaluates model robustness across 11 datasets in RAG, reasoning, alignment, and tool-use tasks against diverse noise types, including random documents, irrelevant chat histories, and hard negative distractors. Our evaluation reveals a catastrophic performance drop of up to 80% in state-of-the-art models when faced with contextual distractors. Crucially, we find that agentic workflows often amplify these errors by over-trusting noisy tool outputs, and distractors can trigger emergent misalignment even without adversarial intent. We find that prompting, context engineering, SFT, and outcome-reward only RL fail to ensure robustness; in contrast, our proposed Rationale-Aware Reward (RARE) significantly strengthens resilience by incentivizing the identification of helpful information within noise. Finally, we uncover an inverse scaling trend where increased test-time computation leads to worse performance in noisy settings and demonstrate via attention visualization that models disproportionately focus on distractor tokens, providing vital insights for building the next generation of robust, reasoning-capable agents.

Janghoon Han, Heegyu Kim, Changho Lee et al.

Recent advances in large language models have enabled deep research systems that generate expert-level reports through multi-step reasoning and evidence-based synthesis. However, evaluating such reports remains challenging: report quality is multifaceted, making it difficult to determine what to assess and by what criteria; LLM-based judges may miss errors that require domain expertise to identify; and because deep research relies on retrieved evidence, report-wide claim verification is also necessary. To address these issues, we propose DEER, a benchmark for evaluating expert-level deep research reports. DEER systematizes evaluation criteria with an expert-developed taxonomy (7 dimensions, 25 subdimensions) operationalized as 101 fine-grained rubric items. We also provide task-specific Expert Evaluation Guidance to support LLM-based judging. Alongside rubric-based assessment, we propose a claim verification architecture that verifies both cited and uncited claims and quantifies evidence quality. Experiments show that while current deep research systems can produce structurally plausible reports that cite external evidence, there is room for improvement in fulfilling expert-level user requests and achieving logical completeness. Beyond simple performance comparisons, DEER makes system strengths and limitations interpretable and provides diagnostic signals for improvement.

Sangwoo Shin, Seunghyun Kim, Youngsoo Jang et al.

In embodied instruction-following (EIF), the integration of pretrained language models (LMs) as task planners emerges as a significant branch, where tasks are planned at the skill level by prompting LMs with pretrained skills and user instructions. However, grounding these pretrained skills in different domains remains challenging due to their intricate entanglement with the domain-specific knowledge. To address this challenge, we present a semantic skill grounding (SemGro) framework that leverages the hierarchical nature of semantic skills. SemGro recognizes the broad spectrum of these skills, ranging from short-horizon low-semantic skills that are universally applicable across domains to long-horizon rich-semantic skills that are highly specialized and tailored for particular domains. The framework employs an iterative skill decomposition approach, starting from the higher levels of semantic skill hierarchy and then moving downwards, so as to ground each planned skill to an executable level within the target domain. To do so, we use the reasoning capabilities of LMs for composing and decomposing semantic skills, as well as their multi-modal extension for assessing the skill feasibility in the target domain. Our experiments in the VirtualHome benchmark show the efficacy of SemGro in 300 cross-domain EIF scenarios.

Hyunseung Lim, Sooyohn Nam, Sungmin Na et al.

Patent examination remains an ongoing challenge in the NLP literature even after the advent of large language models (LLMs), as it requires an extensive yet nuanced human judgment on whether a submitted claim meets the statutory standards of novelty and non-obviousness against previously granted claims -- prior art -- in expert domains. Previous NLP studies have approached this challenge as a prediction task (e.g., forecasting grant outcomes) with high-level proxies such as similarity metrics or classifiers trained on historical labels. However, this approach often overlooks the step-by-step evaluations that examiners must make with profound information, including rationales for the decisions provided in office actions documents, which also makes it harder to measure the current state of techniques in patent review processes. To fill this gap, we construct PANORAMA, a dataset of 8,143 U.S. patent examination records that preserves the full decision trails, including original applications, all cited references, Non-Final Rejections, and Notices of Allowance. Also, PANORAMA decomposes the trails into sequential benchmarks that emulate patent professionals' patent review processes and allow researchers to examine large language models' capabilities at each step of them. Our findings indicate that, although LLMs are relatively effective at retrieving relevant prior art and pinpointing the pertinent paragraphs, they struggle to assess the novelty and non-obviousness of patent claims. We discuss these results and argue that advancing NLP, including LLMs, in the patent domain requires a deeper understanding of real-world patent examination. Our dataset is openly available at https://huggingface.co/datasets/LG-AI-Research/PANORAMA.

Seungone Kim, Juyoung Suk, Ji Yong Cho et al.

As language models (LMs) become capable of handling a wide range of tasks, their evaluation is becoming as challenging as their development. Most generation benchmarks currently assess LMs using abstract evaluation criteria like helpfulness and harmlessness, which often lack the flexibility and granularity of human assessment. Additionally, these benchmarks tend to focus disproportionately on specific capabilities such as instruction following, leading to coverage bias. To overcome these limitations, we introduce the BiGGen Bench, a principled generation benchmark designed to thoroughly evaluate nine distinct capabilities of LMs across 77 diverse tasks. A key feature of the BiGGen Bench is its use of instance-specific evaluation criteria, closely mirroring the nuanced discernment of human evaluation. We apply this benchmark to assess 103 frontier LMs using five evaluator LMs. Our code, data, and evaluation results are all publicly available at https://github.com/prometheus-eval/prometheus-eval/tree/main/BiGGen-Bench.

Muhammad Khalifa, Lajanugen Logeswaran, Moontae Lee et al.

In the context of multi-step reasoning, e.g., with chain-of-thought, language models (LMs) can easily assign a high likelihood to incorrect steps. As a result, decoding strategies that optimize for solution likelihood often yield incorrect solutions. To address this issue, we propose Guiding chain-of-thought ReAsoning with a CorrectnEss Discriminator (GRACE), a stepwise decoding approach that steers the decoding process towards producing correct reasoning steps. GRACE employs a step-level verifier or discriminator trained with a contrastive loss over correct and incorrect steps, which is used during decoding to score next-step candidates based on their correctness. Importantly, GRACE only requires sampling from the LM, without the need for LM training or fine-tuning. Using models from FLAN-T5 and LLaMA families, we evaluate GRACE over four math and two symbolic reasoning tasks, where it exhibits substantial performance gains compared to greedy decoding, verifiers, and self-consistency in most settings. When further combined with self-consistency, GRACE outperforms all the baselines by sizeable margins. Human and LLM evaluations over GSM8K show that GRACE not only improves the final answer accuracy but also the correctness of the intermediate reasoning. Our implementation can be accessed at https://github.com/mukhal/grace.

Yunxiang Zhang, Muhammad Khalifa, Lajanugen Logeswaran et al.

Self-correction has emerged as a promising solution to boost the reasoning performance of large language models (LLMs), where LLMs refine their solutions using self-generated critiques that pinpoint the errors. This work explores whether small (<= 13B) language models (LMs) have the ability of self-correction on reasoning tasks with minimal inputs from stronger LMs. We propose a novel pipeline that prompts smaller LMs to collect self-correction data that supports the training of self-refinement abilities. First, we leverage correct solutions to guide the model in critiquing their incorrect responses. Second, the generated critiques, after filtering, are used for supervised fine-tuning of the self-correcting reasoner through solution refinement. Our experimental results show improved self-correction abilities of two models on five datasets spanning math and commonsense reasoning, with notable performance gains when paired with a strong GPT-4-based verifier, though limitations are identified when using a weak self-verifier for determining when to correct.

Siqi Shen, Lajanugen Logeswaran, Moontae Lee et al.

Large language models (LLMs) have demonstrated substantial commonsense understanding through numerous benchmark evaluations. However, their understanding of cultural commonsense remains largely unexamined. In this paper, we conduct a comprehensive examination of the capabilities and limitations of several state-of-the-art LLMs in the context of cultural commonsense tasks. Using several general and cultural commonsense benchmarks, we find that (1) LLMs have a significant discrepancy in performance when tested on culture-specific commonsense knowledge for different cultures; (2) LLMs' general commonsense capability is affected by cultural context; and (3) The language used to query the LLMs can impact their performance on cultural-related tasks. Our study points to the inherent bias in the cultural understanding of LLMs and provides insights that can help develop culturally aware language models.

Lechen Zhang, Tolga Ergen, Lajanugen Logeswaran et al.

Large Language Models (LLMs) have shown impressive capabilities in many scenarios, but their performance depends, in part, on the choice of prompt. Past research has focused on optimizing prompts specific to a task. However, much less attention has been given to optimizing the general instructions included in a prompt, known as a system prompt. To address this gap, we propose SPRIG, an edit-based genetic algorithm that iteratively constructs prompts from prespecified components to maximize the model's performance in general scenarios. We evaluate the performance of system prompts on a collection of 47 different types of tasks to ensure generalizability. Our study finds that a single optimized system prompt performs on par with task prompts optimized for each individual task. Moreover, combining system and task-level optimizations leads to further improvement, which showcases their complementary nature. Experiments also reveal that the optimized system prompts generalize effectively across model families, parameter sizes, and languages. This study provides insights into the role of system-level instructions in maximizing LLM potential.

Kyungjae Lee, Dasol Hwang, Sunghyun Park et al.

Despite the promise of RLHF in aligning LLMs with human preferences, it often leads to superficial alignment, prioritizing stylistic changes over improving downstream performance of LLMs. Underspecified preferences could obscure directions to align the models. Lacking exploration restricts identification of desirable outputs to improve the models. To overcome these challenges, we propose a novel framework: Reinforcement Learning from Reflective Feedback (RLRF), which leverages fine-grained feedback based on detailed criteria to improve the core capabilities of LLMs. RLRF employs a self-reflection mechanism to systematically explore and refine LLM responses, then fine-tuning the models via a RL algorithm along with promising responses. Our experiments across Just-Eval, Factuality, and Mathematical Reasoning demonstrate the efficacy and transformative potential of RLRF beyond superficial surface-level adjustment.

Sungik Choi, Hankook Lee, Honglak Lee et al.

Novelty detection is a fundamental task of machine learning which aims to detect abnormal ($\textit{i.e.}$ out-of-distribution (OOD)) samples. Since diffusion models have recently emerged as the de facto standard generative framework with surprising generation results, novelty detection via diffusion models has also gained much attention. Recent methods have mainly utilized the reconstruction property of in-distribution samples. However, they often suffer from detecting OOD samples that share similar background information to the in-distribution data. Based on our observation that diffusion models can \emph{project} any sample to an in-distribution sample with similar background information, we propose \emph{Projection Regret (PR)}, an efficient novelty detection method that mitigates the bias of non-semantic information. To be specific, PR computes the perceptual distance between the test image and its diffusion-based projection to detect abnormality. Since the perceptual distance often fails to capture semantic changes when the background information is dominant, we cancel out the background bias by comparing it against recursive projections. Extensive experiments demonstrate that PR outperforms the prior art of generative-model-based novelty detection methods by a significant margin.

Hyungyu Shin, Jingyu Tang, Yoonjoo Lee et al.

Peer review underpins scientific progress, but it is increasingly strained by reviewer shortages and growing workloads. Large Language Models (LLMs) can automatically draft reviews now, but determining whether LLM-generated reviews are trustworthy requires systematic evaluation. Researchers have evaluated LLM reviews at either surface-level (e.g., BLEU and ROUGE) or content-level (e.g., specificity and factual accuracy). Yet it remains uncertain whether LLM-generated reviews attend to the same critical facets that human experts weigh -- the strengths and weaknesses that ultimately drive an accept-or-reject decision. We introduce a focus-level evaluation framework that operationalizes the focus as a normalized distribution of attention across predefined facets in paper reviews. Based on the framework, we developed an automatic focus-level evaluation pipeline based on two sets of facets: target (e.g., problem, method, and experiment) and aspect (e.g., validity, clarity, and novelty), leveraging 676 paper reviews (https://figshare.com/s/d5adf26c802527dd0f62) from OpenReview that consists of 3,657 strengths and weaknesses identified from human experts. The comparison of focus distributions between LLMs and human experts showed that the off-the-shelf LLMs consistently have a more biased focus towards examining technical validity while significantly overlooking novelty assessment when criticizing papers.

Yongrae Jo, Seongyun Lee, Minju Seo et al.

Text-to-SQL models are pivotal for making Electronic Health Records (EHRs) accessible to healthcare professionals without SQL knowledge. With the advancements in large language models, these systems have become more adept at translating complex questions into SQL queries. Nonetheless, the critical need for reliability in healthcare necessitates these models to accurately identify unanswerable questions or uncertain predictions, preventing misinformation. To address this problem, we present a self-training strategy using pseudo-labeled unanswerable questions to enhance the reliability of text-to-SQL models for EHRs. This approach includes a two-stage training process followed by a filtering method based on the token entropy and query execution. Our methodology's effectiveness is validated by our top performance in the EHRSQL 2024 shared task, showcasing the potential to improve healthcare decision-making through more reliable text-to-SQL systems.

Dayoon Ko, Jihyuk Kim, Haeju Park et al.

Large reasoning models (LRMs) have demonstrated strong performance in complex, multi-step reasoning tasks. Existing methods enhance LRMs by sequentially integrating external knowledge retrieval; models iteratively generate queries, retrieve external information, and progressively reason over this information. However, purely sequential querying increases inference latency and context length, diminishing coherence and potentially reducing accuracy. To address these limitations, we introduce HDS-QA (Hybrid Deep Search QA), a synthetic dataset automatically generated from Natural Questions, explicitly designed to train LRMs to distinguish parallelizable from sequential queries. HDS-QA comprises hybrid-hop questions that combine parallelizable independent subqueries (executable simultaneously) and sequentially dependent subqueries (requiring step-by-step resolution), along with synthetic reasoning-querying-retrieval paths involving parallel queries. We fine-tune an LRM using HDS-QA, naming the model HybridDeepSearcher, which outperforms state-of-the-art baselines across multiple benchmarks, notably achieving +15.9 and +11.5 F1 on FanOutQA and a subset of BrowseComp, respectively, both requiring comprehensive and exhaustive search. Experimental results highlight two key advantages: HybridDeepSearcher reaches comparable accuracy with fewer search turns, significantly reducing inference latency, and it effectively scales as more turns are permitted. These results demonstrate the efficiency, scalability, and effectiveness of explicitly training LRMs to leverage hybrid parallel and sequential querying.