A focused feed for RLHF, preference data, rater protocols, agent evaluation, and LLM-as-judge research.
Every paper includes structured metadata for quick triage.
These domains typically involve fixed content and passive consumption, where user preferences can be matched by genre or theme.
In contrast, games present distinct challenges: fast-evolving catalogs, interaction-driven preferences (e.g., skill level, mechanics, hardware), and increased risk of unsafe responses in open-ended conversation.
Recent work suggests that preference-tuning techniques -- such as Reinforcement Learning from Human Feedback (RLHF) methods like PPO and GRPO, as well as alternatives like DPO -- reduce diversity, creating a dilemma given that these models
Using open-ended tasks that require no human intervention, we find counterintuitive results: when using diversity metrics that do not explicitly consider quality, preference-tuned models -- particularly those trained via RL -- often produce
Shangheng Du, Jiabao Zhao, Jinxin Shi, Zhentao Xie, Xin Jiang, Yanhong Bai · Mar 16, 2025
Citations: 0
Simulation EnvLong HorizonGeneral
With the rapid development of Large Language Models (LLMs), LLM-based agents have been widely adopted in various fields, becoming essential for autonomous decision-making and interactive tasks.
However, current work typically relies on prompt design or fine-tuning strategies applied to vanilla LLMs, which often leads to limited effectiveness or suboptimal performance in complex agent-related environments.
Hanjiang Hu, Alexander Robey, Changliu Liu · Feb 28, 2025
Citations: 0
Red TeamAutomatic MetricsGeneral
To address this challenge, we propose a safety steering framework grounded in safe control theory, ensuring invariant safety in multi-turn dialogues.
Our method achieves invariant safety at each turn of dialogue by learning a safety predictor that accounts for adversarial queries, preventing potential context drift toward jailbreaks.
Mike Zhang, Amalie Pernille Dilling, Léon Gondelman, Niels Erik Ruan Lyngdorf, Euan D. Lindsay, Johannes Bjerva · Feb 18, 2025
Citations: 0
Critique EditAutomatic MetricsGeneral
Through comprehensive evaluations with three LLM judges and three human experts, across a subset of 900 outputs, we demonstrate that SEFL-tuned models outperform both their untuned counterparts and an existing baseline in terms of feedback
The potential for societal impact is reinforced by extensive qualitative comments and ratings from human stakeholders -- both students and higher education instructors.
We evaluate the ability of the current generation of large language models (LLMs) to help a decision-making agent facing an exploration-exploitation tradeoff.
Rahul Garg, Trilok Padhi, Hemang Jain, Ugur Kursuncu, Ponnurangam Kumaraguru · Nov 19, 2024
Citations: 0
Automatic MetricsSimulation EnvGeneral
Experimental results from our study on two hate speech benchmark datasets demonstrate superior performance over the state-of-the-art baselines across AU-ROC, F1, and Recall with improvements of 1.1%, 7%, and 35%, respectively.
Safeyah Khaled Alshemali, Daniel Bauer, Yuval Marton · Oct 19, 2024
Citations: 0
Automatic MetricsLong HorizonGeneral
We set a new state-of-the-art on thematic fit benchmarks, but show that closed and open weight LLMs respond differently to our prompting strategies: Closed models achieve better scores overall and benefit from multi-step reasoning, but they
Karen Li, Kopo M. Ramokapane, Awais Rashid · Aug 4, 2022
Citations: 0
Automatic MetricsWeb BrowsingGeneral
Our results uncover various misperceptions about the security controls and how design constraints, e.g., safety and lack of regular updates due to the long-term nature of such systems, provide significant challenges to the realization of mo
