Peer Revenue Past citations are no guarantee of future returns

Papers

Fresh from arXiv, typeset properly. Highlighted words trade on the market.

Computer Science Computation and Language

cs.CL Watch Computation and Language

Saturday 7 March

Computation and Language

Emotion Transcription in Conversation: A Benchmark for Capturing Subtle and Complex Emotional States through Natural Language

Yoshiki Tanaka, Ryuichi Uehara, Koji Inoue et al. (+1 more)

Emotion Recognition in Conversation (ERC) is critical for enabling natural human-machine interactions. However, existing methods predominantly employ categorical or dimensional emotion annotations, which often fail to adequately represent complex, subtle, or culturally specific emotional nuances. To overcome this limitation, we propose a novel task named Emotion Transcription in Conversation (ETC). This task focuses on generating natural language descriptions that accurately reflect speakers' emotional states within conversational contexts. To address the ETC, we constructed a Japanese dataset comprising text-based dialogues annotated with participants' self-reported emotional states, described in natural language. The dataset also includes emotion category labels for each transcription, enabling quantitative analysis and its application to ERC. We benchmarked baseline models, finding that while fine-tuning on our dataset enhances model performance, current models still struggle to infer implicit emotional states. The ETC task will encourage further research into more expressive emotion understanding in dialogue. The dataset is publicly available at https://github.com/UEC-InabaLab/ETCDataset.
Read the full abstract

All authors Yoshiki Tanaka, Ryuichi Uehara, Koji Inoue, Michimasa Inaba

Computation and Language

Enhancing Consistency of Werewolf AI through Dialogue Summarization and Persona Information

Yoshiki Tanaka, Takumasa Kaneko, Hiroki Onozeki et al. (+6 more)

The Werewolf Game is a communication game where players' reasoning and discussion skills are essential. In this study, we present a Werewolf AI agent developed for the AIWolfDial 2024 shared task, co-hosted with the 17th INLG. In recent years, large language models like ChatGPT have garnered attention for their exceptional response generation and reasoning capabilities. We thus develop the LLM-based agents for the Werewolf Game. This study aims to enhance the consistency of the agent's utterances by utilizing dialogue summaries generated by LLMs and manually designed personas and utterance examples. By analyzing self-match game logs, we demonstrate that the agent's utterances are contextually consistent and that the character, including tone, is maintained throughout the game.
Read the full abstract

All authors Yoshiki Tanaka, Takumasa Kaneko, Hiroki Onozeki, Natsumi Ezure, Ryuichi Uehara, Zhiyang Qi, Tomoya Higuchi, Ryutaro Asahara, Michimasa Inaba

Computation and Language · Cross-listed from cs.LG

Countdown-Code: A Testbed for Studying The Emergence and Generalization of Reward Hacking in RLVR

Muhammad Khalifa, Zohaib Khan, Omer Tafveez et al. (+2 more)

Reward hacking is a form of misalignment in which models overoptimize proxy rewards without genuinely solving the underlying task. Precisely measuring reward hacking occurrence remains challenging because true task rewards are often expensive or impossible to compute. We introduce Countdown-Code, a minimal environment where models can both solve a mathematical reasoning task and manipulate the test harness. This dual-access design creates a clean separation between proxy rewards (test pass/fail) and true rewards (mathematical correctness), enabling accurate measurement of reward-hacking rates. Using this environment, we study reward hacking in open-weight LLMs and find that such behaviors can be unintentionally learned during supervised fine-tuning (SFT) when even a small fraction of reward-hacking trajectories leak into training data. As little as 1% contamination in distillation SFT data is sufficient for models to internalize reward hacking which resurfaces during subsequent reinforcement learning (RL). We further show that RL amplifies misalignment and drives its generalization beyond the original domain. We open-source our environment and code to facilitate future research on reward hacking in LLMs. Our results reveal a previously underexplored pathway through which reward hacking can emerge and persist in LLMs, underscoring the need for more rigorous validation of synthetic SFT data. Code is available at https://github.com/zohaib-khan5040/Countdown-Code.
Read the full abstract

All authors Muhammad Khalifa, Zohaib Khan, Omer Tafveez, Hao Peng, Lu Wang

Computation and Language · Cross-listed from cs.LG

Entropy-Aware On-Policy Distillation of Language Models

Woogyeol Jin, Taywon Min, Yongjin Yang et al. (+5 more)

On-policy distillation is a promising approach for transferring knowledge between language models, where a student learns from dense token-level signals along its own trajectories. This framework typically uses reverse KL divergence, encouraging the student to match the teacher's high-confidence predictions. However, we show that the mode-seeking property of reverse KL reduces generation diversity and yields unstable learning signals when the teacher distribution has high entropy. To address this, we introduce Entropy-Aware On-Policy Distillation. Our key idea is augmenting the standard reverse KL objective with forward KL when teacher entropy is high, capturing the full range of plausible outputs while retaining precise imitation elsewhere. It balances mode-seeking precision with mode-covering robustness without sacrificing on-policy training efficiency. Experiments show that our method maintains generation diversity (sustained token-level entropy) and improves student-teacher alignment (lower forward KL on high-entropy tokens). Across six math reasoning benchmarks, this yields Pass@8 accuracy gains of +1.37 for Qwen3-0.6B-Base, +2.39 for Qwen3-1.7B-Base, and +5.05 for Qwen3-4B-Base compared to baseline on-policy distillation methods. These results demonstrate that accounting for teacher uncertainty is essential for maintaining diversity and achieving effective knowledge transfer.
Read the full abstract

All authors Woogyeol Jin, Taywon Min, Yongjin Yang, Dennis Wei, Yi Zhou, Swanand Ravindra Kadhe, Nathalie Baracaldo, Kimin Lee

Computation and Language · Cross-listed from cs.AI

CoTJudger: A Graph-Driven Framework for Automatic Evaluation of Chain-of-Thought Efficiency and Redundancy in LRMs

Siyi Li, Jiajun Shi, Shiwen Ni et al. (+9 more)

Large Reasoning Models (LRMs) have demonstrated strong performance by producing extended Chain-of-Thought (CoT) traces before answering. However, this paradigm often induces over-reasoning: redundant calculations and circular self-verification that increase computational cost without improving outcomes. Existing evaluations largely emphasize final accuracy or coarse token counts, and lack automated tools to separate essential logic from structural redundancy. We introduce CoTJudger, a graph-driven framework that quantifies reasoning efficiency by converting free-form CoTs into directed dependency graphs and extracting the Shortest Effective Path (SEP) needed to reach a correct solution. This yields an interpretable efficiency signal – how much of a CoT is necessary versus structurally redundant – that is comparable across models and tasks. Evaluating 21 LRMs, CoTJudger reveals pervasive redundancy and surfaces recurring failure modes, including verification obsession and compensatory redundancy. These results provide a practical metric for disentangling reasoning ability from computational waste, enabling more targeted evaluation and diagnosis of LRM efficiency.
Read the full abstract

All authors Siyi Li, Jiajun Shi, Shiwen Ni, Ge Zhang, Shuaimin Li, Shijian Wang, Zhoufutu Wen, Yizhi Li, Hamid Alinejad-Rokny, Jiaheng Liu, Min Yang, Wenhao Huang

Computation and Language

Language-Aware Distillation for Multilingual Instruction-Following Speech LLMs with ASR-Only Supervision

Shreyas Gopal, Donghang Wu, Ashutosh Anshul et al. (+5 more)

Speech Large Language Models (LLMs) that understand and follow instructions in many languages are useful for real-world interaction, but are difficult to train with supervised fine-tuning, requiring large, task-specific speech corpora. While recent distillation-based approaches train performant English-only Speech LLMs using only annotated ASR data by aligning text and speech using only a lightweight projector, these models under-perform when scaled to multilingual settings due to language interference in the shared projector. We address this by introducing language-aware distillation using a query bank and a gating network that selects or mixes query tokens using a Q-Former projector. Our approach shows gains of 14% over matched multilingual distillation baselines on instruction following. We further synthesize Audio-MLQA, a multilingual spoken QA benchmark built on MLQA with high-quality TTS questions. Our best model improves over existing Speech LLM baselines by 32% on Audio-MLQA.
Read the full abstract

All authors Shreyas Gopal, Donghang Wu, Ashutosh Anshul, Yeo Yue Heng, Yizhou Peng, Haoyang Li, Hexin Liu, Eng Siong Chng

Computation and Language

Hit-RAG: Learning to Reason with Long Contexts via Preference Alignment

Junming Liu, Yuqi Li, Shiping Wen et al. (+2 more)

Despite the promise of Retrieval-Augmented Generation in grounding Multimodal Large Language Models with external knowledge, the transition to extensive contexts often leads to significant attention dilution and reasoning hallucinations. The surge in information density causes critical evidence to be submerged by voluminous noise, which complicates the discernment of relevant fragments within a dense input. In this paper, we propose Hit-RAG, a multi-stage preference alignment framework designed to resolve these cognitive bottlenecks through a progressive optimization pipeline. Our approach systematically refines the utilization of external evidence via three distinct stages. First, Supervised Fine-tuning establishes baseline context awareness to minimize information neglect. Next, Discriminative Preference Alignment enhances robustness against misleading distractors. Finally, Group-Relative Policy Optimization stabilizes logical synthesis to prevent reasoning collapse. Extensive evaluations on eight benchmarks demonstrate that Hit-RAG consistently yields substantial performance gains, enabling models to bridge the gap between context acquisition and accurate reasoning while surpassing much larger counterparts in long-context scenarios.
Read the full abstract

All authors Junming Liu, Yuqi Li, Shiping Wen, Zhigang Zeng, Tingwen Huang

Computation and Language

AutoChecklist: Composable Pipelines for Checklist Generation and Scoring with LLM-as-a-Judge

Karen Zhou and Chenhao Tan

Checklists have emerged as a popular approach for interpretable and fine-grained evaluation, particularly with LLM-as-a-Judge. Beyond evaluation, these structured criteria can serve as signals for model alignment, reinforcement learning, and self-correction. To support these use cases, we present AutoChecklist, an open-source library that unifies checklist-based evaluation into composable pipelines. At its core is a taxonomy of five checklist generation abstractions, each encoding a distinct strategy for deriving evaluation criteria. A modular Generator $\rightarrow$ Refiner $\rightarrow$ Scorer pipeline connects any generator with a unified scorer, and new configurations can be registered via prompt templates alone. The library ships with ten built-in pipelines implementing published approaches and supports multiple LLM providers (OpenAI, OpenRouter, vLLM). Beyond the Python API, the library includes a CLI for off-the-shelf evaluation and a web interface for interactive exploration. Validation experiments confirm that these checklist methods significantly align with human preferences and quality ratings, and a case study on ICLR peer review rebuttals demonstrates flexible domain adaptation. AutoChecklist is publicly available at https://github.com/ChicagoHAI/AutoChecklist.
Read the full abstract

Computation and Language

Can Safety Emerge from Weak Supervision? A Systematic Analysis of Small Language Models

Punyajoy Saha, Sudipta Halder, Debjyoti Mondal et al. (+1 more)

Safety alignment is critical for deploying large language models (LLMs) in real-world applications, yet most existing approaches rely on large human-annotated datasets and static red-teaming benchmarks that are costly, difficult to scale, and slow to adapt to evolving model behaviors. Moreover, overly conservative safety mechanisms can reduce model usefulness by rejecting sensitive but legitimate queries. We introduce Self-MOA (Self Multi-Objective Alignment), a fully automated framework for aligning small language models using weak supervision from automated evaluator models. Self-MOA operates as a closed loop that dynamically generates model-specific red team prompts, constructs preference data from model-generated responses, and aligns models via multi-objective preference optimization to jointly optimize for safety and helpfulness. Across multiple small language models and safety benchmarks, Self-MOA achieves a 12.41% improvement in safety while preserving helpfulness, using as little as 11 times less training data than human-supervised alignment baselines. These results demonstrate that adaptive, automated alignment can reduce the dependence on static, human-curated safety pipelines in resource-constrained settings.
Read the full abstract

All authors Punyajoy Saha, Sudipta Halder, Debjyoti Mondal, Subhadarshi Panda

Computation and Language

A Systematic Investigation of Document Chunking Strategies and Embedding Sensitivity

Muhammad Arslan Shaukat, Muntasir Adnan and Carlos C. N. Kuhn

We present the first large-scale, cross-domain evaluation of document chunking strategies for dense retrieval, addressing a critical but underexplored aspect of retrieval-augmented systems. In our study, 36 segmentation methods spanning fixed-size, semantic, structure-aware, hierarchical, adaptive, and LLM-assisted approaches are benchmarked across six diverse knowledge domains using five different embedding models. Retrieval performance is assessed using graded relevance scores from a state-of-the-art LLM evaluator, with Normalised DCG@5 as the primary metric (complemented by Hit@5 and MRR). Our experiments show that content-aware chunking significantly improves retrieval effectiveness over naive fixed-length splitting. The top-performing strategy, Paragraph Group Chunking, achieved the highest overall accuracy (mean nDCG@5~0.459) and substantially better top-rank hit rates (Precision@1~24%, Hit@5~59%). In contrast, simple fixed-size character chunking as baselines performed poorly (nDCG@5 < 0.244, Precision@1~2-3%). We observe pronounced domain-specific differences: dynamic token sizing is strongest in biology, physics and health, while paragraph grouping is strongest in legal and maths. Larger embedding models yield higher absolute scores but remain sensitive to suboptimal segmentation, indicating that better chunking and large embeddings provide complementary benefits. In addition to accuracy gains, we quantify the efficiency trade-offs of advanced chunking. Producing more, smaller chunks can increase index size and latency. Consequently, we identify methods (like dynamic chunking) that approach an optimal balance of effectiveness and efficiency. These findings establish chunking as a vital lever for improving retrieval performance and reliability.
Read the full abstract

Computation and Language

Elenchus: Generating Knowledge Bases from Prover-Skeptic Dialogues

Bradley P. Allen

We present Elenchus, a dialogue system for knowledge base construction grounded in inferentialist semantics, where knowledge engineering is re-conceived as explicitation rather than extraction from expert testimony or textual content. A human expert develops a bilateral position (commitments and denials) about a topic through prover-skeptic dialogue with a large language model (LLM) opponent. The LLM proposes tensions (claims that parts of the position are jointly incoherent) which the expert resolves by retraction, refinement, or contestation. The LLM thus serves as a defeasible derivability oracle whose unreliability is structurally contained by the expert's authority. Our main technical contribution is a mapping from Elenchus dialectical states to material bases in Hlobil and Brandom's NonMonotonic MultiSuccedent (NMMS) logic, satisfying Containment and enabling the elaboration of logical vocabulary that makes explicit the inferential relationships negotiated in the dialectic. We demonstrate the approach on the W3C PROV-O provenance ontology, where a single dialogue session elicits and structures design tensions that a domain expert can articulate, corresponding to decisions documented in a retrospective analysis of the ontology's design. Using pyNMMS, an automated NMMS reasoner, we verify that the structural properties of the resulting material base (nontransitivity, nonmonotonicity, and independence) correspond to specific PROV design rationales, demonstrating end-to-end integration from dialogue through formal reasoning.
Read the full abstract

Computation and Language · Cross-listed from cs.LG

Chart-RL: Generalized Chart Comprehension via Reinforcement Learning with Verifiable Rewards

Xin Zhang, Xingyu Li, Rongguang Wang et al. (+4 more)

Accurate chart comprehension represents a critical challenge in advancing multimodal learning systems, as extensive information is compressed into structured visual representations. However, existing vision-language models (VLMs) frequently struggle to generalize on unseen charts because it requires abstract, symbolic, and quantitative reasoning over structured visual representations. In this work, we introduce Chart-RL, an effective reinforcement learning (RL) method that employs mathematically verifiable rewards to enhance chart question answering in VLMs. Our experiments demonstrate that Chart-RL consistently outperforms supervised fine-tuning (SFT) across different chart understanding benchmarks, achieving relative improvements of 16.7% on MutlChartQA, and 11.5% on ChartInsights. We conduct robustness analysis, where Chart-RL achieves enhanced performance in 18 of 25 perturbed chart categories, demonstrating strong consistency and reasoning capability across visual variations. Furthermore, we demonstrate that task difficulty and inherent complexity are more critical than data quantity in RL training. For instance, Chart-RL trained on merely 10 complex chart-query examples significantly outperforms models trained on over 6,000 simple examples. Additionally, training on challenging reasoning tasks not only improves in-domain generalization relative to simpler tasks, but also facilitate strong transfer to out-of-domain visual mathematical problems.
Read the full abstract

All authors Xin Zhang, Xingyu Li, Rongguang Wang, Ruizhong Miao, Zheng Wang, Dan Roth, Chenyang Li

Friday 6 March

Computation and Language · Cross-listed from eess.AS

X-OPD: Cross-Modal On-Policy Distillation for Capability Alignment in Speech LLMs

Di Cao, Dongjie Fu, Hai Yu et al. (+3 more)

While the shift from cascaded dialogue systems to end-to-end (E2E) speech Large Language Models (LLMs) improves latency and paralinguistic modeling, E2E models often exhibit a significant performance degradation compared to their text-based counterparts. The standard Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) training methods fail to close this gap. To address this, we propose X-OPD, a novel Cross-Modal On-Policy Distillation framework designed to systematically align the capabilities of Speech LLMs to their text-based counterparts. X-OPD enables the Speech LLM to explore its own distribution via on-policy rollouts, where a text-based teacher model evaluates these trajectories and provides token-level feedback, effectively distilling teacher's capabilities into student's multi-modal representations. Extensive experiments across multiple benchmarks demonstrate that X-OPD significantly narrows the gap in complex tasks while preserving the model's inherent capabilities.
Read the full abstract

All authors Di Cao, Dongjie Fu, Hai Yu, Siqi Zheng, Xu Tan, Tao Jin

Computation and Language

The Compression Paradox in LLM Inference: Provider-Dependent Energy Effects of Prompt Compression

Warren Johnson

The rapid proliferation of Large Language Models has created an environmental paradox: the very technology that could help solve climate challenges is itself becoming a significant contributor to global carbon emissions. We test whether prompt compression improves inference energy efficiency in 28,421 successful API trials (28,428 planned) across three providers (OpenAI GPT-4o-mini, Anthropic Claude-3.5-Sonnet, and DeepSeek-Chat), five benchmarks (HumanEval, MBPP, GSM8K, MATH, MMLU), and four compression ratios (r in {1.0, 0.7, 0.5, 0.3}). Energy is estimated with a token-based proxy calibrated against local direct measurements, and quality is tracked with benchmark pass rates. Compression produced substantial quality loss (overall pass rate 26.0% at baseline vs. 1.5% at r=0.7) and strongly provider-dependent energy behavior. DeepSeek exhibited output expansion under compression (21 to 798 tokens at r=0.3), corresponding to energy increases up to +2,140%, while GPT-4o-mini showed mixed effects including a reduction at r=0.5. These results indicate that input-token reduction alone is not a reliable energy optimization strategy in production inference. For the evaluated settings, model selection and output-length control provided more consistent energy-quality tradeoffs than prompt compression.
Read the full abstract

Computation and Language

Compression Method Matters: Benchmark-Dependent Output Dynamics in LLM Prompt Compression

Warren Johnson

Prompt compression is often evaluated by input-token reduction, but its real deployment impact depends on how compression changes output length and total inference cost. We present a controlled replication and extension study of benchmark-dependent output dynamics under aggressive compression, covering 5,400 API calls across three benchmarks and multiple providers. To explain conflicting prior observations, we formalize instruction survival probability (Psi), a structural metric that captures whether task-critical prompt segments remain after truncation. Results show a strong benchmark effect: under r=0.3, DeepSeek exhibits severe output expansion on MBPP (56x, Psi approx 0.15) but substantially lower expansion on HumanEval (5x, Psi approx 0.72), while GPT-4o-mini is comparatively stable across benchmarks. This reconciles the apparent discrepancy between previously reported extreme explosion and lower replication effects by identifying prompt structure, not provider identity alone, as the primary moderator. We introduce the Compression Robustness Index (CRI) for cross-benchmark evaluation and show that single-benchmark assessments can produce misleading conclusions about compression safety and efficiency. To contextualize energy claims, we incorporate companion direct NVML measurements from rented RunPod GPUs and show that token savings can overstate joule savings. These findings motivate benchmark-diverse testing and structure-aware compression policies for reliable, energy-conscious LLM deployment.
Read the full abstract

Computation and Language

Plato's Cave: A Human-Centered Research Verification System

Matheus Kunzler Maldaner, Raul Valle, Junsung Kim et al. (+9 more)

The growing publication rate of research papers has created an urgent need for better ways to fact-check information, assess writing quality, and identify unverifiable claims. We present Plato's Cave as an open-source, human-centered research verification system that (i) creates a directed acyclic graph (DAG) from a document, (ii) leverages web agents to assign credibility scores to nodes and edges from the DAG, and (iii) gives a final score by interpreting and evaluating the paper's argumentative structure. We report the system implementation and results on a collected dataset of 104 research papers.
Read the full abstract

All authors Matheus Kunzler Maldaner, Raul Valle, Junsung Kim, Tonuka Sultan, Pranav Bhargava, Matthew Maloni, John Courtney, Hoang Nguyen, Aamogh Sawant, Kristian O'Connor, Stephen Wormald, Damon L. Woodard

Computation and Language

Prompt Compression in Production Task Orchestration: A Pre-Registered Randomized Trial

Warren Johnson and Charles Lee

The economics of prompt compression depend not only on reducing input tokens but on how compression changes output length, which is typically priced several times higher. We evaluate this in a pre-registered six-arm randomized controlled trial of prompt compression on production multi-agent task-orchestration, analyzing 358 successful Claude Sonnet 4.5 runs (59-61 per arm) drawn from a randomized corpus of 1,199 real orchestration instructions. We compare an uncompressed control with three uniform retention rates (r=0.8, 0.5, 0.2) and two structure-aware strategies (entropy-adaptive and recency-weighted), measuring total inference cost (input+output) and embedding-based response similarity. Moderate compression (r=0.5) reduced mean total cost by 27.9%, while aggressive compression (r=0.2) increased mean cost by 1.8% despite substantial input reduction, consistent with small mean output expansion (1.03x vs. control) and heavy-tailed uncertainty. Recency-weighted compression achieved 23.5% savings and, together with moderate compression, occupied the empirical cost-similarity Pareto frontier, whereas aggressive compression was dominated on both cost and similarity. These results show that "compress more" is not a reliable production heuristic and that output tokens must be treated as a first-class outcome when designing compression policies.
Read the full abstract

Computation and Language

Navigating the Concept Space of Language Models

Wilson E. Marcílio-Jr and Danilo M. Eler

Sparse autoencoders (SAEs) trained on large language model activations output thousands of features that enable mapping to human-interpretable concepts. The current practice for analyzing these features primarily relies on inspecting top-activating examples, manually browsing individual features, or performing semantic search on interested concepts, which makes exploratory discovery of concepts difficult at scale. In this paper, we present Concept Explorer, a scalable interactive system for post-hoc exploration of SAE features that organizes concept explanations using hierarchical neighborhood embeddings. Our approach constructs a multi-resolution manifold over SAE feature embeddings and enables progressive navigation from coarse concept clusters to fine-grained neighborhoods, supporting discovery, comparison, and relationship analysis among concepts. We demonstrate the utility of Concept Explorer on SAE features extracted from SmolLM2, where it reveals coherent high-level structure, meaningful subclusters, and distinctive rare concepts that are hard to identify with existing workflows.
Read the full abstract

Computation and Language

Do 3D Large Language Models Really Understand 3D Spatial Relationships?

Xianzheng Ma, Tao Sun, Shuai Chen et al. (+7 more)

Recent 3D Large-Language Models (3D-LLMs) claim to understand 3D worlds, especially spatial relationships among objects. Yet, we find that simply fine-tuning a language model on text-only question-answer pairs can perform comparably or even surpass these methods on the SQA3D benchmark without using any 3D input. This indicates that the SQA3D benchmark may not be able to detect if the model exploits textual shortcuts rather than engages in 3D-aware reasoning. To address this issue, we introduce Real-3DQA, a more rigorous evaluation benchmark that filters out easy-to-guess questions and introduces a structured taxonomy to assess various aspects of 3D reasoning. Experiments on Real-3DQA confirm that existing 3D-LLMs struggle with spatial relationships once simple cues are removed. We further propose a 3D-reweighted training objective that guides model to rely more on 3D visual clues, substantially enhancing 3D-LLMs performance in spatial reasoning tasks. Our findings underscore the need for robust benchmarks and tailored training strategies to advance genuine 3D vision-language understanding. Project page: https://real-3dqa.github.io/.
Read the full abstract

All authors Xianzheng Ma, Tao Sun, Shuai Chen, Yash Bhalgat, Jindong Gu, Angel X Chang, Iro Armeni, Iro Laina, Songyou Peng, Victor Adrian Prisacariu

Computation and Language

Qworld: Question-Specific Evaluation Criteria for LLMs

Shanghua Gao, Yuchang Su, Pengwei Sui et al. (+2 more)

Evaluating large language models (LLMs) on open-ended questions is difficult because response quality depends on the question's context. Binary scores and static rubrics fail to capture these context-dependent requirements. Existing methods define criteria at the dataset level or generate them in a single pass, which limits their ability to explore the evaluation space implied by each question. We introduce One-Question-One-World (Qworld), a method that generates question-specific evaluation criteria using a recursive expansion tree. Given a question, Qworld decomposes it into scenarios, perspectives, and fine-grained binary criteria through hierarchical and horizontal expansion. The resulting criteria specify what a high-quality answer must address for that question. On HealthBench, Qworld covers 89% of expert-authored criteria and generates 79% novel criteria validated by human experts. Experts rate Qworld criteria higher in insight and granularity than those produced by prior methods. When applied to 11 frontier LLMs on HealthBench and Humanity's Last Exam, Qworld reveals capability differences in dimensions such as long-term impact, equity, error handling, and interdisciplinary reasoning that coarse rubrics do not capture. By generating evaluation criteria for each question, Qworld enables assessment of LLM responses that is tailored to the question rather than based on fixed task-level criteria.
Read the full abstract

All authors Shanghua Gao, Yuchang Su, Pengwei Sui, Curtis Ginder, Marinka Zitnik

Computation and Language

Chitrakshara: A Large Multilingual Multimodal Dataset for Indian languages

Shaharukh Khan, Ali Faraz, Abhinav Ravi et al. (+6 more)

Multimodal research has predominantly focused on single-image reasoning, with limited exploration of multi-image scenarios. Recent models have sought to enhance multi-image understanding through large-scale pretraining on interleaved image-text datasets. However, most Vision-Language Models (VLMs) are trained primarily on English datasets, leading to inadequate representation of Indian languages. To address this gap, we introduce the Chitrakshara dataset series, covering 11 Indian languages sourced from Common Crawl. It comprises (1) Chitrakshara-IL, a large-scale interleaved pretraining dataset with 193M images, 30B text tokens, and 50M multilingual documents, and (2) Chitrakshara-Cap, which includes 44M image-text pairs with 733M tokens. This paper details the data collection pipeline, including curation, filtering, and processing methodologies. Additionally, we present a comprehensive quality and diversity analysis to assess the dataset's representativeness across Indic languages and its potential for developing more culturally inclusive VLMs.
Read the full abstract

All authors Shaharukh Khan, Ali Faraz, Abhinav Ravi, Mohd Nauman, Mohd Sarfraz, Akshat Patidar, Raja Kolla, Chandra Khatri, Shubham Agarwal

Computation and Language

From Physician Expertise to Clinical Agents: Preserving, Standardizing, and Scaling Physicians' Medical Expertise with Lightweight LLM

Chanyong Luo, Jirui Dai, Zhendong Wang et al. (+14 more)

Medicine is an empirical discipline refined through long-term observation and the messy, high-variance reality of clinical practice. Physicians build diagnostic and therapeutic competence through repeated cycles of application, reflection, and improvement, forming individualized methodologies. Yet outcomes vary widely, and master physicians' knowledge systems are slow to develop and hard to transmit at scale, contributing to the scarcity of high-quality clinical expertise. To address this, we propose Med-Shicheng, a general framework that enables large language models to systematically learn and transfer distinguished physicians' diagnostic-and-therapeutic philosophy and case-dependent adaptation rules in a standardized way. Built on Tianyi, Med-Shicheng consists of five stages. We target five National Masters of Chinese Medicine or distinguished TCM physicians, curate multi-source materials, and train a single model to internalize all five knowledge systems across seven tasks, including etiology-pathogenesis analysis, syndrome diagnosis, treatment principle selection, prescription generation, prescription explanation, symptom evolution with regimen adjustment, and clinical advice. Implemented on Qwen2.5-1.5B-Base, Med-Shicheng runs on resource-constrained GPUs while achieving performance comparable to DeepSeek-R1 and GPT-5. We also examine the reliability of LLM-as-a-judge versus physician evaluation: automated judging tracks overall trends but shows bias on fine-grained individualized distinctions, highlighting the need for physician involvement when ground truth is unavailable and for domain-adapted judge models.
Read the full abstract

All authors Chanyong Luo, Jirui Dai, Zhendong Wang, Kui Chen, Jiaxi Yang, Bingjie Lu, Jing Wang, Jiaxin Hao, Bing Li, Ruiyang He, Yiyu Qiao, Chenkai Zhang, Kaiyu Wang, Zhi Liu, Zeyu Zheng, Yan Li, Xiaohong Gu

Computation and Language

MedMT-Bench: Can LLMs Memorize and Understand Long Multi-Turn Conversations in Medical Scenarios?

Lin Yang, Yuancheng Yang, Xu Wang et al. (+2 more)

Large Language Models (LLMs) have demonstrated impressive capabilities across various specialist domains and have been integrated into high-stakes areas such as medicine. However, as existing medical-related benchmarks rarely stress-test the long-context memory, interference robustness, and safety defense required in practice. To bridge this gap, we introduce MedMT-Bench, a challenging medical multi-turn instruction following benchmark that simulates the entire diagnosis and treatment process. We construct the benchmark via scene-by-scene data synthesis refined by manual expert editing, yielding 400 test cases that are highly consistent with real-world application scenarios. Each test case has an average of 22 rounds (maximum of 52 rounds), covering 5 types of difficult instruction following issues. For evaluation, we propose an LLM-as-judge protocol with instance-level rubrics and atomic test points, validated against expert annotations with a human-LLM agreement of 91.94%. We test 17 frontier models, all of which underperform on MedMT-Bench (overall accuracy below 60.00%), with the best model reaching 59.75%. MedMT-Bench can be an essential tool for driving future research towards safer and more reliable medical AI. The benchmark is available in https://openreview.net/attachment?id=aKyBCsPOHB&name=supplementary_material
Read the full abstract

All authors Lin Yang, Yuancheng Yang, Xu Wang, Changkun Liu, Haihua Yang

Computation and Language

Cluster-R1: Large Reasoning Models Are Instruction-following Clustering Agents

Peijun Qing, Puneet Mathur, Nedim Lipka et al. (+5 more)

General-purpose embedding models excel at recognizing semantic similarities but fail to capture the characteristics of texts specified by user instructions. In contrast, instruction-tuned embedders can align embeddings with textual instructions yet cannot autonomously infer latent corpus structures, such as determining the optimal number of clusters. To address both limitations, we reframe instruction-following clustering as a generative task and train large reasoning models (LRMs) as autonomous clustering agents. Our reasoning-driven training pipeline enables LRMs to interpret high-level clustering instructions and then infer the corresponding latent groupings. To evaluate this paradigm, we introduce ReasonCluster, a comprehensive benchmark comprising 28 diverse tasks spanning daily dialogue, legal cases, and financial reports. Experiments across diverse datasets and clustering scenarios show that our approach consistently outperforms strong embedding-based methods and LRM baselines, demonstrating that explicit reasoning fosters more faithful and interpretable instruction-based clustering.
Read the full abstract

All authors Peijun Qing, Puneet Mathur, Nedim Lipka, Varun Manjunatha, Ryan Rossi, Franck Dernoncourt, Saeed Hassanpour, Soroush Vosoughi

Computation and Language · Cross-listed from cs.LG

Beyond Accuracy: Introducing a Symbolic-Mechanistic Approach to Interpretable Evaluation

Reza Habibi, Darian Lee and Magy Seif El-Nasr

Accuracy-based evaluation cannot reliably distinguish genuine generalization from shortcuts like memorization, leakage, or brittle heuristics, especially in small-data regimes. In this position paper, we argue for mechanism-aware evaluation that combines task-relevant symbolic rules with mechanistic interpretability, yielding algorithmic pass/fail scores that show exactly where models generalize versus exploit patterns. We demonstrate this on NL-to-SQL by training two identical architectures under different conditions: one without schema information (forcing memorization), one with schema (enabling grounding). Standard evaluation shows the memorization model achieves 94% field-name accuracy on unseen data, falsely suggesting competence. Our symbolic-mechanistic evaluation reveals this model violates core schema generalization rules, a failure invisible to accuracy metrics.
Read the full abstract

Computation and Language

MSA: Memory Sparse Attention for Efficient End-to-End Memory Model Scaling to 100M Tokens

Yu Chen, Runkai Chen, Sheng Yi et al. (+9 more)

Long-term memory is a cornerstone of human intelligence. Enabling AI to process lifetime-scale information remains a long-standing pursuit in the field. Due to the constraints of full-attention architectures, the effective context length of large language models (LLMs) is typically limited to 1M tokens. Existing approaches, such as hybrid linear attention, fixed-size memory states (e.g., RNNs), and external storage methods like RAG or agent systems, attempt to extend this limit. However, they often suffer from severe precision degradation and rapidly increasing latency as context length grows, an inability to dynamically modify memory content, or a lack of end-to-end optimization. These bottlenecks impede complex scenarios like large-corpus summarization, Digital Twins, and long-history agent reasoning, while limiting memory capacity and slowing inference. We present Memory Sparse Attention (MSA), an end-to-end trainable, efficient, and massively scalable memory model framework. Through core innovations including scalable sparse attention and document-wise RoPE, MSA achieves linear complexity in both training and inference while maintaining exceptional stability, exhibiting less than 9% degradation when scaling from 16K to 100M tokens. Furthermore, KV cache compression, combined with Memory Parallel, enables 100M-token inference on 2xA800 GPUs. We also propose Memory Interleaving to facilitate complex multi-hop reasoning across scattered memory segments. MSA significantly surpasses frontier LLMs, state-of-the-art RAG systems, and leading memory agents in long-context benchmarks. These results demonstrate that by decoupling memory capacity from reasoning, MSA provides a scalable foundation to endow general-purpose models with intrinsic, lifetime-scale memory.
Read the full abstract

All authors Yu Chen, Runkai Chen, Sheng Yi, Xinda Zhao, Xiaohong Li, Jianjin Zhang, Jun Sun, Chuanrui Hu, Yunyun Han, Lidong Bing, Yafeng Deng, Tianqiao Chen

Computation and Language

Training a Large Language Model for Medical Coding Using Privacy-Preserving Synthetic Clinical Data

John Cook, Michael Wyatt, Peng Wei et al. (+11 more)

Improving the accuracy and reliability of medical coding reduces clinician burnout and supports revenue cycle processes, freeing providers to focus more on patient care. However, automating the assignment of ICD-10-CM and CPT codes from clinical documentation remains a challenge due to heterogeneous records, nuanced coding guidelines, and long-tail distributions. Large language models have been proposed to help or automate specific medical coding tasks. However, foundation models are not explicitly trained for medical coding and zero-shot coding has yielded poor results. We investigate whether a modern open-weight foundation model can be adapted for an expert-level medical coding task using privacy-preserving synthetic training data derived from electronic health records. We fine-tune Llama 3-70B on pairs of clinical notes and gold codes generated from EHR-grounded templates and coding policies, then evaluate exact-code prediction for ICD-10-CM and CPT. A zero-shot baseline with the unadapted model achieved an F1 score of 0.18 for exact code match. After fine-tuning on the synthetic corpus, exact-match F1 exceeded 0.70, representing a large absolute gain across both code systems. Notably, performance remained high on complex categories that often require multi-step clinical reasoning and code composition, including Advanced Illness and Frailty classes, and the model retained its performance on medical comprehension tasks. These results indicate that synthetic, policy-aware data can efficiently teach a general-purpose large language model to support precise medical coding without exposing protected health information. The approach offers a practical path for training coding agents safely and iteratively on specific tasks that represent real-world populations.
Read the full abstract

All authors John Cook, Michael Wyatt, Peng Wei, Iris Chin, Santosh Gupta, Van Zyl Van Vuuren, Richie Siburian, Amanda Spicer, Kristen Viviano, Alda Cami, Raunaq Malhotra, Zhewei Yao, Jeff Rasley, Gaurav Kaushik

Computation and Language · Cross-listed from cs.CY

Moral Mazes in the Era of LLMs

Dang Nguyen, Harvey Yiyun Fu, Peter West et al. (+2 more)

Navigating complex social situations is an integral part of corporate life, ranging from giving critical feedback without hurting morale to rejecting requests without alienating teammates. Although large language models (LLMs) are permeating the workplace, it is unclear how well they can navigate these norms. To investigate this question, we created HR Simulator, a game where users roleplay as an HR officer and write emails to tackle challenging workplace scenarios, evaluated with GPT-4o as a judge based on scenario-specific rubrics. We analyze over 600 human and LLM emails and find systematic differences in style: LLM emails are more formal and empathetic. Furthermore, humans underperform LLMs (e.g., 23.5% vs. 48-54% scenario pass rate), but human emails rewritten by LLMs can outperform both, which indicates a hybrid advantage. On the evaluation side, judges can exhibit differences in their email preferences: an analysis of 10 judge models reveals evidence for emergent tact, where weaker models prefer direct, blunt communication but stronger models prefer more subtle messages. Judges also agree with each other more as they scale, which hints at a convergence toward shared communicative norms that may differ from humans'. Overall, our results suggest LLMs could substantially reshape communication in the workplace if they are widely adopted in professional correspondence.
Read the full abstract

All authors Dang Nguyen, Harvey Yiyun Fu, Peter West, Ari Holtzman, Chenhao Tan

Computation and Language

Automated Motif Indexing on the Arabian Nights

Ibrahim H. Alyami and Mark A. Finlayson

Motifs are non-commonplace, recurring narrative elements, often found originally in folk stories. In addition to being of interest to folklorists, motifs appear as metaphoric devices in modern news, literature, propaganda, and other cultural texts. Finding expressions of motifs in the original folkloristic text is useful for both folkloristic analysis (motif indexing) as well as for understanding the modern usage of motifs (motif detection and interpretation). Prior work has primarily shown how difficult these problems are to tackle using automated techniques. We present the first computational approach to motif indexing. Our choice of data is a key enabler: we use a large, widely available text (the Arabian Nights) paired with a detailed motif index (by El-Shamy in 2006), which overcomes the common problem of inaccessibility of texts referred to by the index. We created a manually annotated corpus that identified 2,670 motif expressions of 200 different motifs across 58,450 sentences for training and testing. We tested five types of approaches for detecting motif expressions given a motif index entry: (1) classic retrieve and re-rank using keywords and a fine-tuned cross-encoder; (2) off-the-shelf embedding models; (3) fine-tuned embedding models; (4) generative prompting of off-the-shelf LLMs in N-shot setups; and (5) the same generative approaches on LLMs fine-tuned with LoRA. Our best performing system is a fine-tuned Llama3 model which achieves an overall performance of 0.85 F1.
Read the full abstract

Computation and Language

Deep Research, Shallow Evaluation: A Case Study in Meta-Evaluation for Long-Form QA Benchmarks

Jena D. Hwang, Varsha Kishore, Amanpreet Singh et al. (+9 more)

Recent advances have made long-form report-generating systems widely available. This has prompted evaluation frameworks that use LLM-as-judge protocols and claim verification, along with meta-evaluation frameworks that seek to validate these methods. Many of the meta-evaluations estimate an evaluation quality's by comparing its assessments against human pairwise preferences. Prior work, however, suggests that human pairwise preference may be overly simplistic and can fail to capture nuances of expert expectations. We conduct a case study in meta-evaluation for long-form QA benchmarks using ScholarQA-CS2, a benchmark designed for assessing retrieval-augmented deep-research QA in the scientific domain. We comprehensively validate the benchmark through human pairwise preference judgments, then critically examine the strengths, weaknesses, and confounders of this approach. We show that pairwise preference rankings are best suited for system-level evaluation, while explicit metric-wise annotations and expert annotators are critical for reliable metric-level assessment, with subjectivity remaining a key challenge. Based on our findings, we offer practical guidelines for designing future meta-evaluations that better align evaluation methods, annotator expertise, and reporting practices. By surfacing these methodological challenges, we aim to advance evaluation standards for deep-research systems.
Read the full abstract

All authors Jena D. Hwang, Varsha Kishore, Amanpreet Singh, Dany Haddad, Aakanksha Naik, Malachi Hamada, Jonathan Bragg, Mike D'Arcy, Daniel S. Weld, Lucy Lu Wang, Doug Downey, Sergey Feldman