Mourad Heddaya

Reasoning language models generate long reasoning traces that increase latency and cost. We study how to shorten these traces while preserving accuracy on competition-level mathematics. We compare three approaches in a teacher-student distillation setup: (i) inference-time truncation after the first k tokens; (ii) Implicit Chain-of-Thought (ICoT)-style curricula that progressively shorten the teacher trace during training; and (iii) direct distillation to shorter reasoning traces. We find: (1) with standard SFT and first-k truncation, models compensate by generating longer text after reasoning, undermining token savings; (2) ICoT-style curricula provide little benefit on competition-level mathematics with long, diverse reasoning traces; and (3) training on post-think (text the teacher generates after reasoning) outperforms generic summaries by 4–5 percentage points at matched token budgets. These results show that curriculum-based methods that are effective on simple tasks do not transfer to complex reasoning, while post-think provides a better distillation target because it preserves the teacher's solution path.

We conduct a large-scale, century-spanning analysis of U.S. inflation narratives by applying natural language processing methods to American newspapers published from 1923 to 2025. We analyze 4.2 million sentences, documenting significant shifts in narrative prevalence across economic eras. We find that fiscal explanations dominated pre-1980 discussion, while monetary and supply-side factors gained prominence thereafter. We also find sharp regional differences, with Northern papers emphasizing fiscal causes while Southern papers focus on supply factors and interest rate effects. We examine narrative diffusion, finding evidence that newspapers engage in narrative differentiation from local competitors while simultaneously experiencing contagion effects from more distant sources. Narrative framing also predicts heterogeneity in both short- and long-term consumer inflation expectations across income and education groups, with these effects in some cases exceeding the predictive power of realized inflation itself.

Bilateral negotiation is a complex, context-sensitive task in which human negotiators dynamically adjust anchors, pacing, and flexibility to exploit power asymmetries and informal cues. We introduce a unified mathematical framework for modeling concession dynamics based on a hyperbolic tangent curve, and propose two metrics—burstiness and the Concession-Rigidity Index (CRI)—to quantify the timing and rigidity of offer trajectories. We conduct a large-scale empirical comparison between human negotiators and four state-of-the-art large language models across natural-language and numeric-offers settings. Our results reveal that, unlike humans who smoothly adapt to situations and infer the opponent's position and strategies, LLMs systematically anchor at extremes of the possible agreement zone and optimize for fixed points irrespective of leverage or context. These findings highlight fundamental limitations in current LLM negotiation capabilities and point to the need for models that better internalize opponent reasoning and context-dependent strategy.

This paper introduces CaseSumm, a novel dataset for long-context summarization in the legal domain that addresses the need for longer and more complex datasets for summarization evaluation. We collect 25.6K U.S. Supreme Court (SCOTUS) opinions and their official summaries, known as "syllabuses." Our dataset is the largest open legal case summarization dataset, and is the first to include summaries of SCOTUS decisions dating back to 1815.
We also present a comprehensive evaluation of LLM-generated summaries using both automatic metrics and expert human evaluation, revealing discrepancies between these assessment methods. Our evaluation shows Mistral 7b, a smaller open-source model, outperforms larger models on most automatic metrics and successfully generates syllabus-like summaries. In contrast, human expert annotators indicate that Mistral summaries contain hallucinations. The annotators consistently rank GPT-4 summaries as clearer and exhibiting greater sensitivity and specificity. Further, we find that LLM-based evaluations are not more correlated with human evaluations than traditional automatic metrics. Furthermore, our analysis identifies specific hallucinations in generated summaries, including precedent citation errors and misrepresentations of case facts. These findings demonstrate the limitations of current automatic evaluation methods for legal summarization and highlight the critical role of human evaluation in assessing summary quality, particularly in complex, high-stakes domains.

We present a novel approach to classify causal micro-narratives from text. These narratives are sentence-level explanations of the cause(s) and/or effect(s) of a target subject. The approach requires only a subject-specific ontology of causes and effects, and we demonstrate it with an application to inflation narratives. Using a human-annotated dataset spanning historical and contemporary US news articles for training, we evaluate several large language models (LLMs) on this multi-label classification task. The best-performing model---a fine-tuned Llama 3.1 8B---achieves F1 scores of 0.87 on narrative detection and 0.71 on narrative classification. Comprehensive error analysis reveals challenges arising from linguistic ambiguity and highlights how model errors often mirror human annotator disagreements. This research establishes a framework for extracting causal micro-narratives from real-world data, with wide-ranging applications to social science research.

Leveraging an established exercise in negotiation education, we build a novel dataset for studying how the use of language shapes bilateral bargaining. Our dataset extends existing work in two ways: 1) we recruit participants via behavioral labs instead of crowdsourcing platforms and allow participants to negotiate through audio, enabling more naturalistic interactions; 2) we add a control setting where participants negotiate only through alternating, written numeric offers. Despite the two contrasting forms of communication, we find that the average agreed prices of the two treatments are identical. But when subjects can talk, fewer offers are exchanged, negotiations finish faster, the likelihood of reaching agreement rises, and the variance of prices at which subjects agree drops substantially. We further propose a taxonomy of speech acts in negotiation and enrich the dataset with annotated speech acts. We set up prediction tasks to predict negotiation success and find that being reactive to the arguments of the other party is advantageous over driving the negotiation.