Social media platforms have become primary sources for news consumption due to their real-time and interactive nature, yet they have also facilitated the widespread proliferation of misinformation, negatively impacting public health, social cohesion, and market stability. While professional fact-checking is essential for debunking rumors, the process is time-consuming, necessitating automation to effectively combat fake news. Existing approaches, such as extractive methods, often lack coherence and context, whereas abstractive methods leveraging large language models (LLMs) can generate more readable and informative debunking passages. However, readability alone is insufficient for effective misinformation correction; user acceptance is critical. Recent advancements in LLMs offer new opportunities for personalized debunking, as these models can generate context-sensitive responses and adapt to user profiles. Building on this, we propose the MUti-round Refinement and Simulated fEedback-enhanced framework (MURSE), which generates Chinese user-specific debunking passages by iteratively refining outputs based on simulated user feedback. Specifically, MURSE-generated user-specific debunking passages were preferred twice as often as general debunking passages in most cases, highlighting its potential to improve misinformation correction and foster positive dissemination chains.

Reward modeling is a key step in building safe foundation models when applying reinforcement learning from human feedback (RLHF) to align Large Language Models (LLMs). However, reward modeling based on the Bradley-Terry (BT) model assumes a global reward function, failing to capture the inherently diverse and heterogeneous human preferences. Hence, such oversimplification limits LLMs from supporting personalization and pluralistic alignment. Theoretically, we show that when human preferences follow a mixture distribution of diverse subgroups, a single BT model has an irreducible error. While existing solutions, such as fine-grained annotations via prompting or structured preference elicitation, help address this issue, they are costly and constrained by predefined attributes, failing to fully capture the richness of human values. In this work, we introduce MiCRo, a two-stage framework that enhances personalized preference learning by leveraging large-scale binary preference datasets without requiring explicit fine-grained annotations. In the first stage, MiCRo employs a mixture of preferences to model diverse human preferences, enabling a flexible representation of diverse value systems. In the second stage, MiCRo integrates an online routing strategy that dynamically adapts mixture weights based on specific context to resolve ambiguity, allowing for efficient and scalable preference adaptation with minimal additional supervision. Experiments on multiple preference datasets demonstrate that MiCRo effectively captures diverse human preferences and significantly improves personalized preference learning on downstream tasks.

Recent advancements in large language models (LLMs) have spurred growing interest in automatic theorem proving using Lean4, where effective tree search methods are crucial for navigating the underlying large proof search spaces. While the existing approaches primarily rely on value functions and/or Monte Carlo Tree Search (MCTS), the potential of simpler methods like Best-First Tree Search (BFS) remains underexplored. In this paper, we investigate whether BFS can achieve competitive performance in large-scale theorem proving tasks. We present BFS-Prover, a scalable expert iteration framework, featuring three key innovations. First, we implement strategic data filtering at each expert iteration round, excluding problems solvable via beam search node expansion to focus on harder cases. Second, we improve the sample efficiency of BFS through Direct Preference Optimization (DPO) applied to state-tactic pairs automatically annotated with compiler error feedback, refining the LLM’s policy to prioritize productive expansions. Third, we employ length normalization in BFS to encourage exploration of deeper proof paths. BFS-Prover achieves a state-of-the-art score of 72.95 on the MiniF2F test set and therefore challenges the perceived necessity of complex tree search methods, demonstrating that BFS can achieve competitive performance when properly scaled.

COVID-19 has disproportionately threatened minority communities in the U.S, not only in health but also in societal impact. However, social scientists and policymakers lack critical data to capture the dynamics of the anti-Asian hate trend and to evaluate its scale and scope. We introduce new datasets from Twitter related to anti-Asian hate sentiment before and during the pandemic. Relying on Twitter’s academic API, we retrieve hateful and counter-hate tweets from the Twitter Historical Database. To build contextual understanding and collect related racial cues, we also collect instances of heated arguments, often political, but not necessarily hateful, discussing Chinese issues. We then use the state-of-the-art hate speech classifiers to discern whether these tweets express hatred. These datasets can be used to study hate speech, general anti-Asian or Chinese sentiment, and hate linguistics by social scientists as well as to evaluate and build hate speech or sentiment analysis classifiers by computational scholars.