A focused feed for RLHF, preference data, rater protocols, agent evaluation, and LLM-as-judge research.
Every paper includes structured metadata for quick triage.
Ran Xu, Jingjing Chen, Jiayu Ye, Yu Wu, Jun Yan, Carl Yang · Oct 27, 2025
Citations: 0
Pairwise PreferenceHuman EvalCoding
Large Language Models (LLMs) are widely used as judges to evaluate response quality, providing a scalable alternative to human evaluation.
However, most LLM judges operate solely on intrinsic text-based reasoning, limiting their ability to verify complex constraints or perform accurate computation.
Using the best performing LLM as the backbone of a quantitative study with 41 participants, we uncover distinct user preferences across hinting strategies, and identify the limitations of automatic evaluation metrics to capture them.
Ruiyi Yang, Hao Xue, Imran Razzak, Hakim Hacid, Flora D. Salim · Oct 23, 2025
Citations: 0
Automatic MetricsLong HorizonGeneral
A Head Agent provides guidance that leads retrieval, while an Iteration Agent selects and expands HSeq via structure-respecting actions (e.g., parent/child hops, table row/column neighbors, KG relations); Finally the head agent composes can
Experiments on HotpotQA (text), HybridQA/TAT-QA (table+text), and MetaQA (KG) show consistent EM/F1 gains over strong single-pass, multi-hop, and agentic RAG baselines with high efficiency.
Aligning large language models with human preferences is critical for creating reliable and controllable AI systems.
A human preference can be visualized as a high-dimensional vector where different directions represent trade-offs between desired attributes (e.g., helpfulness vs.
Amith Ananthram, Elias Stengel-Eskin, Lorena A. Bradford, Julia Demarest, Adam Purvis, Keith Krut · Oct 21, 2025
Citations: 0
Rubric RatingHuman EvalLlm As JudgeGeneral
While vision-language models (VLMs) have advanced into detailed image description, evaluation remains a challenge.
In this work, we introduce PoSh, a metric for detailed image description that uses scene graphs as structured rubrics to guide LLMs-as-a-Judge, producing aggregate scores grounded in fine-grained errors (e.g.
Chengshu Li, Mengdi Xu, Arpit Bahety, Hang Yin, Yunfan Jiang, Huang Huang · Oct 21, 2025
Citations: 0
DemonstrationsSimulation EnvLong HorizonGeneral
Imitation learning from large-scale, diverse human demonstrations has been shown to be effective for training robots, but collecting such data is costly and time-consuming.
This challenge intensifies for multi-step bimanual mobile manipulation, where humans must teleoperate both the mobile base and two high-DoF arms.
Wei-Jer Chang, Akshay Rangesh, Kevin Joseph, Matthew Strong, Masayoshi Tomizuka, Yihan Hu · Oct 20, 2025
Citations: 0
DemonstrationsSimulation EnvMulti AgentGeneral
Developing autonomous vehicles (AVs) requires not only safety and efficiency, but also realistic, human-like behaviors that are socially aware and predictable.
Achieving this requires sim agent policies that are human-like, fast, and scalable in multi-agent settings.
Marisa C. Peczuh, Nischal Ashok Kumar, Ryan Baker, Blair Lehman, Danielle Eisenberg, Caitlin Mills · Oct 14, 2025
Citations: 0
Rubric RatingAutomatic MetricsGeneral
As the world becomes increasingly saturated with AI-generated content, disinformation, and algorithmic persuasion, critical thinking - the capacity to evaluate evidence, detect unreliable claims, and exercise independent judgment - is becom
We developed a coding rubric based on an established skills progression and completed human coding for a corpus of student essays.
Experiments across multiple domains demonstrate that PRoH achieves state-of-the-art performance, surpassing the prior SOTA model HyperGraphRAG by an average of 19.73% in F1 and 8.41% in Generation Evaluation (G-E) score, while maintaining s
Rongzhi Zhang, Liqin Ye, Yuzhao Heng, Xiang Chen, Tong Yu, Lingkai Kong · Oct 14, 2025
Citations: 0
Pairwise PreferenceAutomatic MetricsCoding
Finally, we demonstrate efficiency enhancements across three downstream tasks: preference data synthesis, Pareto frontier approximation and optimization, and distillation of aligned behaviors for intervention-free inference.
Donggyu Lee, Hyeok Yun, Meeyoung Cha, Sungwon Park, Sangyoon Park, Jihee Kim · Oct 8, 2025
Citations: 0
Automatic MetricsSimulation EnvGeneral
To address this, we introduce EconCausal, a large-scale benchmark comprising 10,490 context-annotated causal triplets extracted from 2,595 high-quality empirical studies published in top-tier economics and finance journals.
Our evaluation reveals critical limitations in current LLMs' context-dependent reasoning.
Xinyu Guo, Zhengliang Shi, Minglai Yang, Mahdi Rahimi, Mihai Surdeanu · Oct 7, 2025
Citations: 0
Human EvalAutomatic MetricsGeneral
Finally, human evaluation shows that our best model generates relational keywords closely aligned with gold labels, increasing human explanation quality ratings by 54% (relative).
Yao Xiao, Jung-jae Kim, Roy Ka-wei Lee, Lidong Bing · Oct 7, 2025
Citations: 0
Pairwise PreferenceAutomatic MetricsGeneral
Self-play preference optimization has emerged as a prominent paradigm for aligning large language models (LLMs).
It typically involves a language model to generate on-policy responses for prompts and a reward model (RM) to guide the selection of chosen and rejected responses, which can be further trained with direct preference optimization (DPO).
Punya Syon Pandey, Hai Son Le, Devansh Bhardwaj, Rada Mihalcea, Zhijing Jin · Oct 6, 2025
Citations: 0
Critique EditAutomatic MetricsGeneral
Yet, existing safety benchmarks rarely test vulnerabilities in domains such as political manipulation, propaganda and disinformation generation, or surveillance and information control.
Our evaluations reveal several shortcomings: open-weight models exhibit high vulnerability to harmful compliance, with Mistral-7B reaching attack success rates as high as 97% to 98% in domains such as historical revisionism, propaganda, and
Zixin Song, Bowen Zhang, Qian-Wen Zhang, Di Yin, Xing Sun, Chunping Li · Oct 5, 2025
Citations: 0
Pairwise PreferenceAutomatic MetricsCoding
On the official C-STS benchmark, PoLi-RL achieves a Spearman correlation coefficient of 48.18, establishing a new SOTA for the cross-encoder architecture.
These inconsistencies stem from a lack of controlled evaluation frameworks and analysis of models' internals to isolate when and why modality interactions support or undermine reasoning.
We address this gap through a logic-grounded evaluation framework that categorizes multimodal reasoning into six interaction patterns, varying how facts are distributed across modalities and logically combined.
Haoyu He, Haozheng Luo, Yan Chen, Qi R. Wang · Sep 27, 2025
Citations: 0
Automatic MetricsLong HorizonCoding
Predicting human mobility is inherently challenging due to complex long-range dependencies and multi-scale periodic behaviors.
To address this, we introduce RHYTHM (Reasoning with Hierarchical Temporal Tokenization for Human Mobility), a unified framework that leverages large language models (LLMs) as general-purpose spatio-temporal predictors and trajectory reason
Optimistic exploration is central to improving sample efficiency in reinforcement learning with human feedback, yet existing exploratory bonus methods to incentivize exploration often fail to realize optimism.
Jiho Kim, Junseong Choi, Woosog Chay, Daeun Kyung, Yeonsu Kwon, Yohan Jo · Sep 26, 2025
Citations: 0
Pairwise PreferenceSimulation EnvGeneral
In our simulation environment, a user agent with a rich persona interacts with the assistant, providing ratings on how well each suggestion aligns with its preferences and context.
Built on ProPerSim, we propose ProPerAssistant, a retrieval-augmented, preference-aligned assistant that continually learns and adapts through user feedback.
