25 papers
Audio-Zero: Label-Free Self-Evolution for Fine-Grained Audio Reasoning
Siqian Tong, Xuan Li, Chaozhuo Li +5
Large Audio Language models (LALMs) have made rapid progress on acoustic understanding, yet they still struggle with fine-grained audio reasoning (e.g., recognizing event order, re…
Multimodal Reward Hacking in Reinforcement Learning
Jiayu Yao, Yiwei Wang, Anmeng Zhang +5
Reinforcement learning (RL) is increasingly used to align multimodal large language models (MLLMs), but higher rewards do not always imply better task performance. This risk is amp…
Lost in a Single Vector: Improving Long-Document Retrieval with Chunk Evidence Aggregation
Shanshan Lyu, Yiwei Wang, Yujun Cai +2
Dense retrieval ranks one query vector against one document vector. On long documents, this interface can fail when a short but decisive span is weakened during document encoding b…
SAW: Stage-Aware Dynamic Weighting for Multi-Objective Reinforcement Learning in Large Language Models
Yuchen He, Baolong Bi, Shenghua Liu +7
Although multi-objective reinforcement learning (MORL) is central to aligning large language models with complex human preferences, the prevailing practice of static weighted summa…
Readable Yet Unpredictable: Rotated-Outcome Prediction in Vision-Language Models
Lexin Wang, Shenghua Liu, Yiwei Wang +2
Can vision-language models predict what a 180° rotation would reveal from the original image alone? We study this ability through Rotated-Outcome Prediction: given an original ima…
Beyond Black-Box Interventions: Latent Probing for Faithful Retrieval-Augmented Generation
Linfeng Gao, Qinggang Zhang, Baolong Bi +9
Retrieval-Augmented Generation (RAG) systems often fail to maintain contextual faithfulness, generating responses that conflict with the provided context or fail to fully leverage…