Context utilization, the ability of Language Models (LMs) to incorporate relevant information from the provided context when generating responses, remains largely opaque to users, who cannot determine whether models draw from parametric memory or provided context, nor identify which specific context pieces inform the response. Highlight explanations (HEs) offer a natural solution as they can point to the exact context pieces and tokens that influenced model outputs. However, no existing work evaluates their effectiveness in accurately explaining context utilization. We address this gap by introducing the first gold standard HE evaluation framework for context attribution, using controlled test cases with known ground-truth context usage, thereby avoiding the limitations of existing indirect proxy evaluations. To demonstrate the framework’s broad applicability, we evaluate four HE methods—three established techniques and MechLight, a mechanistic interpretability approach we adapt for this task—across four context scenarios, four datasets, and five LMs. Overall, we find that MechLight performs best across all context scenarios. However, our findings reveal systematic failures in all methods: Explanation accuracy degrades significantly with context length, and all methods exhibit strong positional biases in multi-document settings. Surprisingly, widely used gradient-based methods provide little value for understanding context usage. These results challenge HEs’ utility in retrieval-augmented generation, factual verification, and other applications. Our framework provides the foundation for developing accurate context attribution methods.

La technologie ASR excelle dans la transcription précise des discours lus préparés, mais elle rencontre encore des défis lorsqu’il s’agit de conversations spontanées. Cela est en partie dû au fait que ces dernières relèvent d’un registre de langage informel, avec disfluences et réductions de parole. Afin de mieux comprendre les différences de production en fonction des styles de parole, nous présentons la création d’un corpus de parole conversationnelle, dont des extraits sont ensuite lus par leurs auteurs. Le corpus comprend 36 heures de parole en chinois mandarin avec leur transcription, réparties entre conversations spontanées et lecture. Nous avons utilisé WHISPER pour la transcription automatique de la parole et le Montreal Forced Aligner pour l’alignement forcé, résultant dans un corpus de parole transcrit avec annotations multi-niveaux incluant phonèmes, caractères/syllabes et mots. De telles productions de parole parallèles (en modes spontané et lu) seront particulièrement intéressantes pour l’étude des réductions temporelles.

This paper presents the winning system that participated in SemEval-2021 Task 5: Toxic Spans Detection. This task aims to locate those spans that attribute to the text’s toxicity within a text, which is crucial for semi-automated moderation in online discussions. We formalize this task as the Sequence Labeling (SL) problem and the Span Boundary Detection (SBD) problem separately and employ three state-of-the-art models. Next, we integrate predictions of these models to produce a more credible and complement result. Our system achieves a char-level score of 70.83%, ranking 1/91. In addition, we also explore the lexicon-based method, which is strongly interpretable and flexible in practice.

Argument pair extraction (APE) aims to extract interactive argument pairs from two passages of a discussion. Previous work studied this task in the context of peer review and rebuttal, and decomposed it into a sequence labeling task and a sentence relation classification task. However, despite the promising performance, such an approach obtains the argument pairs implicitly by the two decomposed tasks, lacking explicitly modeling of the argument-level interactions between argument pairs. In this paper, we tackle the APE task by a mutual guidance framework, which could utilize the information of an argument in one passage to guide the identification of arguments that can form pairs with it in another passage. In this manner, two passages can mutually guide each other in the process of APE. Furthermore, we propose an inter-sentence relation graph to effectively model the inter-relations between two sentences and thus facilitates the extraction of argument pairs. Our proposed method can better represent the holistic argument-level semantics and thus explicitly capture the complex correlations between argument pairs. Experimental results show that our approach significantly outperforms the current state-of-the-art model.