Luca Mouchel

Luca Mouchel, Thiemo Wambsganss, Paola Mejia-Domenzain et al.

Revision behavior in adaptive writing support systems is an important and relatively new area of research that can improve the design and effectiveness of these tools, and promote students' self-regulated learning (SRL). Understanding how these tools are used is key to improving them to better support learners in their writing and learning processes. In this paper, we present a novel pipeline with insights into the revision behavior of students at scale. We leverage a data set of two groups using an adaptive writing support tool in an educational setting. With our novel pipeline, we show that the tool was effective in promoting revision among the learners. Depending on the writing feedback, we were able to analyze different strategies of learners when revising their texts, we found that users of the exemplary case improved over time and that females tend to be more efficient. Our research contributes a pipeline for measuring SRL behaviors at scale in writing tasks (i.e., engagement or revision behavior) and informs the design of future adaptive writing support systems for education, with the goal of enhancing their effectiveness in supporting student writing. The source code is available at https://github.com/lucamouchel/Understanding-Revision-Behavior.

Luca Mouchel, Debjit Paul, Shaobo Cui et al.

Despite the remarkable performance of Large Language Models (LLMs) in natural language processing tasks, they still struggle with generating logically sound arguments, resulting in potential risks such as spreading misinformation. To address this issue, we introduce FIPO, a fallacy-informed framework that leverages preference optimization methods to steer LLMs toward logically sound arguments. FIPO includes a classification loss, to capture the fine-grained information on fallacy types. Our results on argumentation datasets show that our method reduces the fallacy errors by up to 17.5%. Furthermore, our human evaluation results indicate that the quality of the generated arguments by our method significantly outperforms the fine-tuned baselines, as well as other preference optimization methods, such as DPO. These findings highlight the importance of ensuring models are aware of logical fallacies for effective argument generation. Our code is available at github.com/lucamouchel/Logical-Fallacies.

Luca Mouchel, Pierre Bouquet, Yossi Sheffi

This position paper argues that job exposure to AI should be measured with grounded, evidence-based methods, not inferred from LLM priors alone. Current theoretical exposure measures use zero-shot prompting to classify task-level AI exposure, generating labels with no explicit evidence, no transparent chain of reasoning, and no external validation. The stakes of these measurements are too high to rely on such methods, as they influence policy making, where public and private funds are directed, and how workers understand their future prospects. We therefore argue that AI capability claims should meet three standards: reproducibility, external grounding, and inspectability. We propose a retrieval-augmented framework that assigns AI exposure labels to all 18,796 occupation--task pairs in O*NET 30.2, using open-weight reasoning and instruct models with retrieved news articles and academic paper abstracts as evidence of current AI capabilities. Relative to a zero-shot baseline, the grounded condition is preferred in over 72\% of disagreement cases under both automatic and human evaluation, and yields scores that align more closely with observed real-world AI usage. Taken together, these findings suggest that evidence-grounded measurement better captures what current AI systems can plausibly do in practice, rather than what a model asserts without external evidence. Because AI capabilities continue to change, the measurements used to inform policy must evolve with them: theoretical AI exposure scores should be periodically reassessed, not inherited as immutable ground truth.

Shaobo Cui, Junyou Li, Luca Mouchel et al.

To address this gap, our study introduces the concept of causal epistemic consistency, which focuses on the self-consistency of Large Language Models (LLMs) in differentiating intermediates with nuanced differences in causal reasoning. We propose a suite of novel metrics -- intensity ranking concordance, cross-group position agreement, and intra-group clustering -- to evaluate LLMs on this front. Through extensive empirical studies on 21 high-profile LLMs, including GPT-4, Claude3, and LLaMA3-70B, we have favoring evidence that current models struggle to maintain epistemic consistency in identifying the polarity and intensity of intermediates in causal reasoning. Additionally, we explore the potential of using internal token probabilities as an auxiliary tool to maintain causal epistemic consistency. In summary, our study bridges a critical gap in AI research by investigating the self-consistency over fine-grained intermediates involved in causal reasoning.

Alejandro Hernández-Cano, Alexander Hägele, Allen Hao Huang et al. · eth-zurich

We present Apertus, a fully open suite of large language models (LLMs) designed to address two systemic shortcomings in today's open model ecosystem: data compliance and multilingual representation. Unlike many prior models that release weights without reproducible data pipelines or regard for content-owner rights, Apertus models are pretrained exclusively on openly available data, retroactively respecting robots.txt exclusions and filtering for non-permissive, toxic, and personally identifiable content. To mitigate risks of memorization, we adopt the Goldfish objective during pretraining, strongly suppressing verbatim recall of data while retaining downstream task performance. The Apertus models also expand multilingual coverage, training on 15T tokens from over 1800 languages, with ~40% of pretraining data allocated to non-English content. Released at 8B and 70B scales, Apertus approaches state-of-the-art results among fully open models on multilingual benchmarks, rivalling or surpassing open-weight counterparts. Beyond model weights, we release all scientific artifacts from our development cycle with a permissive license, including data preparation scripts, checkpoints, evaluation suites, and training code, enabling transparent audit and extension.