From the 1 of 11 linked papers with an AI index.
11 papers
Divergence Decoding: Training-Free Capability Fusion
Yimi Wang, Hao Li, Shuo Yang +6
The paper proposes Divergence Decoding, a training‑free method that dynamically routes token generation between a generalist LLM and a domain‑specialist LLM using Jensen‑Shannon di…
AsFT: Anchoring Safety During LLM Fine-Tuning Within Narrow Safety Basin
Shuo Yang, Qihui Zhang, Yuyang Liu +7
Fine-tuning large language models (LLMs) improves performance but introduces critical safety vulnerabilities: even minimal harmful data can severely compromise safety measures. We…
Clipping Bottleneck: Stabilizing RLVR via Stochastic Recovery of Near-Boundary Signals
Shuo Yang, Jinda Lu, Chiyu Ma +8
Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as a central paradigm for scaling LLM reasoning, yet its optimization often suffers from training instability and…
One-Way Policy Optimization for Self-Evolving LLMs
Shuo Yang, Jinda Lu, Kexin Huang +6
Reinforcement Learning with Verifiable Rewards (RLVR) has become a promising paradigm for scaling reasoning capabilities of Large Language Models (LLMs). However, the sparsity of b…
Reasoning Portability: Guiding Continual Learning for MLLMs in the RLVR Era
Qiuhe Hong, Yuyang Liu, Shuo Yang +3
Vision-Language Models in Continual Learning (VLM-CL) aim to continuously adapt to new multimodal tasks while retaining prior knowledge. The emerging paradigm that couples Multimod…
CoT-lized Diffusion: Let's Reinforce T2I Generation Step-by-step
Zheyuan Liu, Munan Ning, Qihui Zhang +8
Current text-to-image (T2I) generation models struggle to align spatial composition with the input text, especially in complex scenes. Even layout-based approaches yield suboptimal…