HFEPX Daily Papers for 2026-05-28

Researcher Quick Triage

Use this archive page for time-slice monitoring (what changed in evaluation methods, metrics, and protocol quality this period). Quality band: High .

• 3 papers report explicit quality controls for this archive period.
• Prioritize papers with both benchmark and metric anchors for reliable longitudinal comparisons.

Primary action: Use this slice for trend comparison: review top papers first, then validate shifts in the protocol matrix.

Why This Time Slice Matters

• Use this archive slice to monitor protocol drift and shifts in evaluation methods over 2026-05-28.

Protocol Takeaways For This Period

• Evaluation modes for this slice cluster around automatic_metrics, simulation_env.

Start Here (Highest-Signal Papers In This Slice)

Ranked by protocol completeness and evidence density for faster period-over-period review.

• Recovering Diversity Without Losing Alignment: A DPO Recipe for Post-Trained LLMs

  May 28, 2026 · Citations: 0 · Score: 7.5

  Eval: Automatic Metrics · Metrics: Success rate, Jailbreak success rate

• Latent Performance Profiling of Large Language Models

  May 28, 2026 · Citations: 0 · Score: 6.5

  Eval: Automatic Metrics · Metrics: Accuracy

• Does The Way You Plan Matter? An Empirical Study of Planning Representations for LLM Web Agents

  May 28, 2026 · Citations: 0 · Score: 6.5

  Eval: Simulation Env · Metrics: Task success

• MedCase-Structured: A Text-to-FHIR Dataset for Benchmarking Diagnostic Reasoning in Clinically Realistic EHR Settings

  May 28, 2026 · Citations: 0 · Score: 6.0

  Eval: Automatic Metrics · Metrics: Accuracy

• Token-Level Generalization in LoRA Adapter Backdoors: Attack Characterization and Behavioral Detection

  May 28, 2026 · Citations: 0 · Score: 6.0

  Eval: Automatic Metrics · Metrics: Accuracy, Recall

• SEAL: Can Saturated Benchmarks Be Revived by LLM-as-a-Meta-Judge?

  May 28, 2026 · Citations: 0 · Score: 6.0

  Eval: Automatic Metrics · Metrics: Accuracy, Spearman

Protocol Matrix (Top 10)

Quickly compare method ingredients across this archive slice.

Papers In This Archive Slice

• LLMSurgeon: Diagnosing Data Mixture of Large Language Models

  Yaxin Luo, Jiacheng Cui, Xiaohan Zhao, Xinyi Shang, Jiacheng Liu · May 28, 2026 · Citations: 0

  To evaluate, we introduce LLMScan, a recipe-verifiable evaluation suite built from open-source LLMs with transparent pretraining mixtures.

• SchGen: PCB Schematic Generation with Semantic-Grounded Code Representations

  Qinpei Luo, Ruichun Ma, Xinyu Zhang, Lili Qiu · May 28, 2026 · Citations: 0

  We further construct a large-scale dataset of PCB schematics paired with user prompts via a human-agent collaborative pipeline that converts open-source hardware designs into our representation.

• Unlocking the Working Memory of Large Language Models for Latent Reasoning

  Lukas Aichberger, Sepp Hochreiter · May 28, 2026 · Citations: 0

  In contrast, human cognition can use working memory to hold and manipulate information internally without the need to externalize intermediate thoughts.

• Locally Coherent, Globally Incoherent: Bounding Compositional Incoherence in Multi-Component LLM Agents

  Anany Kotawala · May 28, 2026 · Citations: 0

  Multi-component LLM agents assemble probabilistic claims from components that each see only part of a joint problem; the composition can violate basic probability axioms even when every component is locally coherent.

• Demystifying Data Organization for Enhanced LLM Training

  Yalun Dai, Yangyu Huang, Tongshen Yang, Yonghan Wang, Xin Zhang · May 28, 2026 · Citations: 0

  Abstract shows limited direct human-feedback or evaluation-protocol detail; use as adjacent methodological context.

• COMPOSE: Composing Future Theorems from Citations and Formal Structure

  David Busbib, Michael Werman · May 28, 2026 · Citations: 0

  To support this setting, we construct a dataset of 108K paired scientific-formal graph examples from arXiv and Mathlib, together with a benchmark of 47K future papers from 2024--2025.

• Reasoning with Sampling: Cutting at Decision Points

  Felix Zhou, Anay Mehrotra, Quanquan C. Liu · May 28, 2026 · Citations: 0

  Across MATH500, HumanEval, GPQA Diamond, and AIME26, our method consistently improves over baselines and RL-trained models.

• On Language Generation in the Limit with Bounded Memory

  Jon Kleinberg, Anay Mehrotra, Amin Saberi, Grigoris Velegkas · May 28, 2026 · Citations: 0

  Abstract shows limited direct human-feedback or evaluation-protocol detail; use as adjacent methodological context.

• Resolution Diagnostics for Paired LLM Evaluation

  Anany Kotawala · May 28, 2026 · Citations: 0

  Pairwise Preference

  Across two public LLM leaderboards, many displayed pairwise rankings do not meet a conventional paired-test resolution target under the actual paired evaluation design: 11 of 40 Open LLM Leaderboard v1 pairwise comparisons and 4 of 9…

• MedCase-Structured: A Text-to-FHIR Dataset for Benchmarking Diagnostic Reasoning in Clinically Realistic EHR Settings

  Valentina Bui Muti, Eugénie Dulout, Ziquan Fu · May 28, 2026 · Citations: 0

  Expert Verification

  We introduce a pipeline for generating clinically realistic HL7 FHIR R4 bundles from unstructured text, enabling controllable evaluation of clinical decision support systems.

• Self-Trained Verification for Training- and Test-Time Self-Improvement

  Chen Henry Wu, Aditi Raghunathan · May 28, 2026 · Citations: 0

  Abstract shows limited direct human-feedback or evaluation-protocol detail; use as adjacent methodological context.

• Qwen-VLA: Unifying Vision-Language-Action Modeling across Tasks, Environments, and Robot Embodiments

  Qiuyue Wang, Mingsheng Li, Jian Guan, Jinhui Ye, Sicheng Xie · May 28, 2026 · Citations: 0

  Demonstrations Long Horizon

  Qwen-VLA is trained with a large-scale joint pretraining recipe over diverse data sources, including robotics manipulation trajectories, human egocentric demonstrations, synthetic simulation data, vision-and-language navigation data,…

• Loong: A Human-Like Long Document Translation Agent with Observe-and-Act Adaptive Context Selection

  Yutong Wang, Xuebo Liu, Derek F. Wong, Zhilin Li, Rongqing Jiang · May 28, 2026 · Citations: 0

  Pairwise Preference

  To address this, we propose a human-like long document translation agent called Loong, which leverages a 3E memory module (Essence-Exemplar-Entity) to store summaries, sentence pairs, and entity records as historical context.

• LLUMI: Improving LLM Writing Assistance for Mental Health Support with Online Community Feedback

  Jiwon Kim, Maya Ajit, Sherry Gong, Soorya Ram Shimgekar, Dong Whi Yoo · May 28, 2026 · Citations: 0

  Pairwise Preference

  Large language models (LLMs) show promise in generating supportive responses for mental health queries, but improving their usefulness, empathy, and safety often requires substantial compute, expert input, and labeled data.

• LoMo: Local Modality Substitution for Deeper Vision-Language Fusion

  Feng Han, Zhixiong Zhang, Zheming Liang, Yibin Wang, Jiaqi Wang · May 28, 2026 · Citations: 0

  Pairwise Preference

  Such data bias leads VLMs to exhibit distinct preferences for information acquisition across different modalities.

• How LoRA Remembers? A Parametric Memory Law for LLM Finetuning

  Ziwen Xu, Haiwen Hong, Linsong Yu, Benglei Cui, Longtao Huang · May 28, 2026 · Citations: 0

  While Low-Rank Adaptation (LoRA) is widely used for such memory updates, existing studies mainly rely on qualitative downstream evaluations, leaving the quantitative capacity limits and underlying dynamics of exact parametric memory largely…

• VideoFDB: Evaluating Full-Duplex Vision-Speech Capabilities in Conversational Agents

  Amrita Mazumdar, Seonwook Park, Rajarshi Roy, Nikhil Srihari, Shengze Wang · May 28, 2026 · Citations: 0

  Rubric Rating

  In this work, we present VideoFDB, the first benchmark to evaluate full-duplex audio-visual-to-audio-visual (AV2AV) conversational agents.

• Same Evidence, Different Answers: Canonical-Context On-Policy Distillation for Multi-Turn Language Models

  Zizhuo Lin, Quanling Liu, Jinsheng Quan, Chao Zhang, Yifan Zhu · May 28, 2026 · Citations: 0

  Abstract shows limited direct human-feedback or evaluation-protocol detail; use as adjacent methodological context.

• Knowing What to Solve Before How: Preplan Empowered LLM Mathematical Reasoning

  Shaojie Wang, Liang Zhang · May 28, 2026 · Citations: 0

  Experiments across four backbones and five mathematical reasoning benchmarks show that PPC achieves the best results on 39 of 40 metrics, improving maj@16 and pass@16 by +2.23 and +3.06 over the strongest baseline without introducing…

• CommunityFact: A Dynamic, Multilingual, Multi-domain Benchmark for Misinformation Detection in the Wild

  Sahajpreet Singh, Insyirah Mujtahid, Min-Yen Kan, Kokil Jaidka · May 28, 2026 · Citations: 0

  Misinformation verification increasingly occurs in public, fast-moving, and multilingual online settings, where static benchmarks provide an incomplete measure of model reliability.

• GRASP: Plan-Guided Graph Retrieval with Adaptive Fusion and Reranking on Semi-Structured Knowledge Bases

  Yicheng Tao, Yiqun Wang, Xiangchen Song, Xin Luo, Kai Liu · May 28, 2026 · Citations: 0

  GRASP substantially advances the state of the art on every metric across the three STaRK benchmarks, lifting average Hit@1 from 62.0 to 73.9.

• Do Language Models Track Entities Across State Changes?

  Zilu Tang, Qiao Zhao, Gabriel Franco, Derry Wijaya, Aaron Mueller · May 28, 2026 · Citations: 0

  Behavioral results inform mechanistic hypotheses, and insights from mechanistic analyses help build stronger behavioral evaluations by predicting failure modes missing from existing evaluations.

• How's it going? Reinforcement learning in language models recruits a functional welfare axis

  Andy Q Han, David J. Chalmers, Pavel Izmailov · May 28, 2026 · Citations: 0

  Demonstrations

  We present evidence that RL recruits a pre-existing representation of functional welfare: an estimate of how well or badly the system is doing, relative to its goals.

• When Should Models Change Their Minds? Contextual Belief Management in Large Language Models

  Haoming Xu, Weihong Xu, Zongrui Li, Mengru Wang, Yunzhi Yao · May 28, 2026 · Citations: 0

  Long Horizon

  To make CBM measurable, we introduce BeliefTrack, a closed-world benchmark spanning Rule Discovery and Circuit Diagnosis, where a finite belief space and symbolic verifiers enable exact turn-level evaluation.

• GRUFF: LLM Pronoun Fidelity, Reasoning, and Biases in German

  Fabian Mewes, Anne Lauscher, Vagrant Gautam · May 28, 2026 · Citations: 0

  Abstract shows limited direct human-feedback or evaluation-protocol detail; use as adjacent methodological context.

• A Dual-Path Architecture for Scaling Compute and Capacity in LLMs

  Markus Frey, Behzad Shomali, Joachim Koehler, Mehdi Ali · May 28, 2026 · Citations: 0

  We show that across two FLOP budgets, our dual-path model surpasses iso-FLOP matched models on language modeling and downstream evaluations, while using fewer parameters than the baseline at matched FLOPs.

• Token-Level Generalization in LoRA Adapter Backdoors: Attack Characterization and Behavioral Detection

  Travis Lelle · May 28, 2026 · Citations: 0

  Abstract shows limited direct human-feedback or evaluation-protocol detail; use as adjacent methodological context.

• Do Proactive Agents Really Need an LLM to Decide When to Wake and What to Anchor?

  Xiaoze Liu, Ruowang Zhang, Amir H. Abdi, Michel Galley, Zhikai Chen · May 28, 2026 · Citations: 0

  Proactive agents read user activity as text and call an LLM on every event to decide whether to act.

• CorPipe at CRAC 2026: Empty Nodes and Cross-Lingual Transfer in Multilingual Coreference Resolution

  Milan Straka · May 28, 2026 · Citations: 0

  Furthermore, we perform a series of ablation experiments with different model sizes, empty node prediction methods, and cross-lingual zero-shot evaluation.

• CCS: Clinical Consensus Selection for Radiology Report Generation

  Xi Zhang, Yingshu Li, Zaiqiao Meng, Jake Lever, Edmond S. L. Ho · May 28, 2026 · Citations: 0

  Abstract shows limited direct human-feedback or evaluation-protocol detail; use as adjacent methodological context.

• PARCEL: Pool-Anchored Resampling with Conditioned Elastic Queries for Efficient Vision-Language Understanding

  Selim Kuzucu, Alessio Tonioni, Vasile Lup, Bernt Schiele, Federico Tombari · May 28, 2026 · Citations: 0

  Extensive evaluations across 27 benchmarks show that PARCEL improves the performance-efficiency Pareto frontier, consistently outperforming existing matryoshka baselines across visual-token budgets while preserving the "train once, deploy…

• Dial HEALTHDIAL for Advice: A Multilingual and Multi-Parallel Spoken Dialogue Dataset for Knowledge-Grounded Information Seeking

  Songbo Hu, Yinhong Liu, Ej Zhou, Evgeniia Razumovskaia, Xiaobin Wang · May 28, 2026 · Citations: 0

  We report benchmark results across key dialogue tasks, which reveal consistent performance disparities across languages, even among high-resource ones.

• SEAL: Can Saturated Benchmarks Be Revived by LLM-as-a-Meta-Judge?

  Jiamin Chen, Yidi Wu, Qiexiang Wang, Qianben Chen, Yuchen Li · May 28, 2026 · Citations: 0

  Pairwise Preference

  Therefore, we present Seeded Elimination with Adaptive LLM-as-a-Meta-Judge, a self-improving evaluation protocol for extracting latent ranking signal from saturated benchmarks.

• DirectorBench: Diagnosing Long-Form Video Generation with Personalized Multi-Agent Evaluation

  Jiamin Chen, Qianben Chen, Jiawen Zhang, Yidi Wu, Yuchen Li · May 28, 2026 · Citations: 0

  Pairwise Preference Multi Agent

  However, evaluating such videos remains challenging, since existing benchmarks largely focus on local visual quality, short-horizon temporal consistency, or generic prompt alignment, and provide limited diagnosis of workflow failures and…

• Conformal Certification of Reasoning Trace Prefixes

  Matt Y. Cheung, Ashok Veeraraghavan, Hanjie Chen, Guha Balakrishnan · May 28, 2026 · Citations: 0

  Abstract shows limited direct human-feedback or evaluation-protocol detail; use as adjacent methodological context.

• Adaptive Targeted Dynamic Chunking for Tokenization-Free Hierarchical Model

  Thang Dang, Akira Nakagawa, Kenichi Kobayashi, Koichi Shirahata · May 28, 2026 · Citations: 0

  Evaluations conducted on the FineWeb-Edu 100B dataset demonstrate that hierarchical models equipped with ATDC achieve competitive Bits-Per-Byte (BPB) performance compared to conventional baselines operating at both byte and token levels.

• UniSteer: Text-Guided Flow Matching in Activation Space for Versatile LLM Steering

  Yingdong Shi, Ruiming Zhang, Changming Li, Zhiyu Yang, Kaixing Zhang · May 28, 2026 · Citations: 0

  Abstract shows limited direct human-feedback or evaluation-protocol detail; use as adjacent methodological context.

• HEART-Bench: Do LLM Agents Exhibit Human-like Psychology?

  Weihan Peng, Chenxu Zhang, Qianao Wang, Yuling Shi, Heng Lian · May 28, 2026 · Citations: 0

  While LLM agents have demonstrated remarkable task-oriented abilities such as planning, reasoning, and action, few works have treated them as complete human personalities where emotional dimensions hold equal importance.

• REPOT: Recoverable Program-of-Thought via Checkpoint Repair

  Parsa Mazaheri · May 28, 2026 · Citations: 0

  Long Horizon

  On Derail-550, our controlled recovery benchmark, every condition with access to checkpoint information clears >=30% on GPT-medium and >=70% on Gemini, vs <=3.1% for error-only feedback -- showing that checkpoint information, not the…

• Who Am I? History-Aware Profiles for Student Simulation in Tutoring Dialogues

  Zhangqi Duan, Shuyan Huang, Alexander Scarlatos, Jaewook Lee, Simon Woodhead · May 28, 2026 · Citations: 0

  A key part of developing large language model (LLM)-powered, automated tutoring tools is student simulation, i.e., using LLMs to role-play as students, which can facilitate tutor model evaluation and training.

• Token Inflation: How Dishonest Providers Can Overcharge for Large Language Model Usage

  Shahinul Hoque, Jinghuai Zhang, Jinyuan Sun, Fnu Suya · May 28, 2026 · Citations: 0

  Red Team

  We show that this kind of billing is hard to audit by design: providers hide the model, the tokenizer, and the execution to protect their IP, mitigate jailbreaks, and preserve user privacy, which means an auditor can only inspect proofs the…

• Teaching Values to Machines: Simulating Human-Like Behavior in LLMs

  Asaf Yehudai, Naama Rozen, Ariel Gera · May 28, 2026 · Citations: 0

  Large Language Models (LLMs) demonstrate a remarkable capacity to adopt different personas and roles; however, it remains unclear whether they can manifest behavior that adheres to a coherent, human-like value structure.

• Audio Jailbreaks in Large Audio-Language Models: Taxonomy, Attack-Defense Analysis, and Cost-Aware Evaluation

  Bo-Han Feng, Yu-Hsuan Li Liang, Chien-Feng Liu, You-Hsuan Chang, Yun-Nung Chen · May 28, 2026 · Citations: 0

  Red Team

  Existing work studies these risks under heterogeneous threat models and evaluation protocols, making it difficult to compare attack practicality or defense utility.

• Give it Space! Explicit Disentangling of Positional and Semantic Representations in Encoders

  Pierre-Antoine Lequeu, Camille Barboule, Benjamin Piwowarski · May 28, 2026 · Citations: 0

  The disentangled approach preserves positional encoding, which improves linguistic representation on 49 of the 65 linguistic phenomena of the Flash-Holmes probing benchmark.

• Recovering Diversity Without Losing Alignment: A DPO Recipe for Post-Trained LLMs

  Vinay Samuel, Yapei Chang, Mohit Iyyer · May 28, 2026 · Citations: 0

  Pairwise Preference

  For each prompt, REDIPO samples responses from both base and instruct models, rewrites base-model responses with the instruct model, filters candidates for safety and instruction-following quality, and builds preference pairs that favor…

• Latent Performance Profiling of Large Language Models

  Tanmoy Chakraborty, Ayan Sengupta, Suparna Bhattacharya, Partha Pratim Chakrabarti, Amlan Chakrabarti · May 28, 2026 · Citations: 0

  Large language models (LLMs) frequently achieve impressive scores on standardized benchmarks, yet accuracy alone offers a limited view of their capabilities.

• Adapting Multilingual Embedding Models to Turkish via Cross-Lingual Tokenizer Surgery and Offline Distillation

  M. Ali Bayram, Banu Diri, Savaş Yıldırım · May 28, 2026 · Citations: 0

  Abstract shows limited direct human-feedback or evaluation-protocol detail; use as adjacent methodological context.

• MIC: Maximizing Informational Capacity in Adaptive Representations via Isotropic Subspace Alignment

  Dang Hong Nguyen, Nhi Ngoc-Yen Nguyen, Huy-Hieu Pham · May 28, 2026 · Citations: 0

  Abstract shows limited direct human-feedback or evaluation-protocol detail; use as adjacent methodological context.

• Causal Interventions on Continuous Variables: A Case Study on Verb Bias in Steering Vectors for In-Context Learning

  Zhenghao Herbert Zhou, R. Thomas McCoy, Robert Frank · May 28, 2026 · Citations: 0

  Pairwise Preference

  We show that verb bias is causally represented in steering vectors extracted from large language models: counterfactual edits to verb bias systematically shift downstream structural preferences.

• MuPHI: Learning Implicit Multimodal Harm Reasoning via Semantically Grounded Reward Optimization

  Anisha Saha, Varsha Suresh, Teodora Kamova, Sophia Wiedmann, Timothy Hospedales · May 28, 2026 · Citations: 0

  Our findings suggest that reasoning-oriented reward optimization offers a promising direction towards building multimodal systems that generalize beyond benchmark-specific shortcuts.

• Does The Way You Plan Matter? An Empirical Study of Planning Representations for LLM Web Agents

  Alejandra Zambrano, Sara Vera Marjanovic, Imene Kerboua, Xing Han Lù, Leila Kosseim · May 28, 2026 · Citations: 0

  To address this, we introduce PlanAhead, a static planner-executor framework that evaluates the impact of plan representation in agent performance.

• ExCAM: Explainable Cultural Awareness Metrics

  Christoph Leiter, Haiyue Song, Hour Kaing, Jin Tei, Hideki Tanaka · May 28, 2026 · Citations: 0

  To address this gap, we introduce ExCAM, an Explainable Cultural Awareness Metric, which is, to our knowledge, the first dedicated evaluation metric that identifies, rates and explains cultural errors in instruction-output pairs.

• Internal Representation, Not Clinical Knowledge: Where Apparent LLM Triage Failures Originate

  David Fraile Navarro, Berardino Como, Jialei Sheng, Soundariya Ananthan, Shlomo Berkovsky · May 28, 2026 · Citations: 0

  Patient-voiced clinical-triage benchmarks report high under-triage rates for consumer LLMs for constrained multiple-choice output, yet the same cases score differently with free-text.

• CRITIC-R1: Learning Structured Critics for Retrieval-Augmented Generation

  Wenhan Xiao, Ziwei Zhang, Chuanyue Yu, Xingcheng Fu, Qingyun Sun · May 28, 2026 · Citations: 0

  Critique Edit

  To learn these capabilities, we design two reward functions: Conservative Judgement Alignment (CJA) first encourages calibrated high-level judgements while mitigating the over-aggressive phenomenon, whereas Diagnostic Quality Alignment…

• Towards Verifiable Multimodal Deep Research: A Multi-Agent Harness for Interleaved Report Generation

  Chenghao Zhang, Guanting Dong, Yufan Liu, Tong Zhao, Zhicheng Dou · May 28, 2026 · Citations: 0

  Tool Use

  Large Language Models (LLMs) have advanced autonomous agents from deep search, which retrieves concise factual answers, to deep research, which synthesizes scattered evidence into long-form reports.

• MELD: Mel-Spectrogram-Based Speech Language Modeling with Discrete Latent Variables

  Sung-Lin Yeh, Wei Zhou, Gil Keren, Duc Le, Zhong Meng · May 28, 2026 · Citations: 0

  Abstract shows limited direct human-feedback or evaluation-protocol detail; use as adjacent methodological context.

• EvoRubric: Self-Evolving Rubric-Driven RL for Open-Ended Generation

  Xin Guan, Xiaomeng Hu, Shen Huang, Zhenyi Wang, Bo Zhang · May 28, 2026 · Citations: 0

  Rubric Rating

  Current rubric-based RL methods mitigate this by employing explicit criteria; however, they rely heavily on static, human-annotated rubrics that inevitably cause policy lag, or expensive external proprietary models for dynamic updates.

• Towards Localized and Disentangled Knowledge Editing for Multimodal Large Language Models

  Leijiang Gu, Zhen Zeng, Feng Li, Xinjian Gao, Zenglin Shi · May 28, 2026 · Citations: 0

  Extensive experiments across various benchmarks and MLLMs demonstrate that LDKE achieves superior performance in propagating edits to related contexts while maintaining high locality.

• PRAIB: Peer Review AI Benchmark of Behaviour of LLM-Assisted Reviewing

  Krzysztof Żurawicki, Julia Farganus, Arkadiusz Gaweł, Mateusz Bystroński, Tomasz Jan Kajdanowicz · May 28, 2026 · Citations: 0

  Yet, it remains unknown whether LLMs engage with scientific manuscripts in the same manner as human reviewers, or whether they merely produce review-looking text.

• Data filtering methods for training language models

  Egor Shevchenko, Elena Bruches · May 28, 2026 · Citations: 0

  Label errors, present even in widely used benchmarks, introduce noise into training data and reduce model generalization.

Recent Archive Slices