Yuechi Zhou

Peng Xia, Yuechi Zhou, Ziyan Zhang et al.

In view of the poor robustness of existing Chinese grammatical error correction models on attack test sets and large model parameters, this paper uses the method of knowledge distillation to compress model parameters and improve the anti-attack ability of the model. In terms of data, the attack test set is constructed by integrating the disturbance into the standard evaluation data set, and the model robustness is evaluated by the attack test set. The experimental results show that the distilled small model can ensure the performance and improve the training speed under the condition of reducing the number of model parameters, and achieve the optimal effect on the attack test set, and the robustness is significantly improved. Code is available at https://github.com/Richard88888/KD-CGEC.

Xiaocheng Luo, Kang Wang, Zaifu Zhan et al.

The Chain-of-Thought (CoT) paradigm, while enhancing the interpretability of Large Language Models (LLMs), is constrained by the inefficiencies and expressive limits of natural language. Latent Chain-of-Thought (latent CoT) reasoning, which operates in a continuous latent space, offers a promising alternative but faces challenges from structural complexities in existing multi-step or multi-model paradigms, such as error propagation and coordination overhead. In this paper, we introduce One-Model One-Step, a novel compression framework for Latent Reasoning with Rule-Based Priors(RuPLaR) to address this challenge. Our method trains an LLM to autonomously generate latent reasoning tokens in a single training stage, guided by rule-based prior probability distributions, thereby eliminating cascaded processes and inter-model dependencies. To ensure reasoning quality, we design a joint training objective that enforces answer consistency via cross-entropy, aligns soft tokens with rule-based priors via KL divergence (the Soft Thinking constraint), and adds a problem-thought semantic alignment constraint in the representation space. Extensive experiments show that our compression framework not only improves accuracy by 11.1% over existing latent CoT methods but also achieves this with minimal token usage, underscoring its effectiveness and extensibility. Code: https://github.com/xiaocen-luo/RuPLaR.

Yi Su, Zhenxu Tian, Dan Qiao et al.

Recent reasoning models such as OpenAI-o1 and DeepSeek-R1 have shown strong performance on complex tasks including mathematical reasoning and code generation. However, this performance gain comes with substantially longer output sequences, leading to significantly increased deployment costs. In particular, long outputs require large KV caches, resulting in high memory consumption and severe bandwidth pressure during attention computation. Most existing KV cache optimization methods are designed for long-input, short-output scenarios and are ineffective for the long-output setting of reasoning models. Moreover, importance estimation in prior work is computationally expensive and becomes prohibitive when continuous re-evaluation is required during long generation. To address these challenges, we propose LongFlow, a KV cache compression method with an efficient importance estimation metric derived from an intermediate result of attention computation using only the current query. This design introduces negligible computational overhead and requires no auxiliary storage. We further develop a custom kernel that fuses FlashAttention, importance estimation, and token eviction into a single optimized operator, improving system-level efficiency. Experiments show that LongFlow achieves up to an 11.8 times throughput improvement with 80% KV cache compression with minimal impact on model accuracy.

Dan Qiao, Yi Su, Pinzheng Wang et al.

Large Language Models (LLMs) have played an important role in many fields due to their powerful capabilities.However, their massive number of parameters leads to high deployment requirements and incurs significant inference costs, which impedes their practical applications. Training smaller models is an effective way to address this problem. Therefore, we introduce OpenBA-V2, a 3.4B model derived from multi-stage compression and continual pre-training from the original 15B OpenBA model. OpenBA-V2 utilizes more data, more flexible training objectives, and techniques such as layer pruning, neural pruning, and vocabulary pruning to achieve a compression rate of 77.3\% with minimal performance loss. OpenBA-V2 demonstrates competitive performance compared to other open-source models of similar size, achieving results close to or on par with the 15B OpenBA model in downstream tasks such as common sense reasoning and Named Entity Recognition (NER). OpenBA-V2 illustrates that LLMs can be compressed into smaller ones with minimal performance loss by employing advanced training objectives and data strategies, which may help deploy LLMs in resource-limited scenarios.

Yi Su, Quantong Qiu, Yuechi Zhou et al.

Large Language Models (LLMs) require substantial computational resources during generation. While the Key-Value (KV) cache significantly accelerates this process by storing attention intermediates, its memory footprint grows linearly with sequence length, batch size, and model size, creating a bottleneck in long-context scenarios. Various KV cache compression techniques, including token eviction, quantization, and low-rank projection, have been proposed to mitigate this bottleneck, often complementing each other. This paper focuses on enhancing token eviction strategies. Token eviction leverages the observation that the attention patterns are often sparse, allowing for the removal of less critical KV entries to save memory. However, this reduction usually comes at the cost of notable accuracy degradation, particularly under high compression ratios. To address this issue, we propose \textbf{CaliDrop}, a novel strategy that enhances token eviction through calibration. Our preliminary experiments show that queries at nearby positions exhibit high similarity. Building on this observation, CaliDrop performs speculative calibration on the discarded tokens to mitigate the accuracy loss caused by token eviction. Extensive experiments demonstrate that CaliDrop significantly improves the accuracy of existing token eviction methods.