Hybrid Reinforcement: When Reward Is Sparse, It's Better to Be Dense
Paper • 2510.07242 • Published • 30
Hengxiang Zhang
hxiang
AI & ML interests
None yet
Recent Activity
updated a model about 1 month ago
hxiang/olmo3-7b-GRPO-step-450 published a model about 1 month ago
hxiang/olmo3-7b-GRPO-step-450 liked a Space about 1 month ago
SUSTech/ChineseSafe-BenchmarkOrganizations
upvoted a paper 11 months ago
upvoted 3
papers 12 months ago
Understanding Language Prior of LVLMs by Contrasting Chain-of-Embedding
Paper • 2509.23050 • Published • 15
Clean First, Align Later: Benchmarking Preference Data Cleaning for Reliable LLM Alignment
Paper • 2509.23564 • Published • 8
LUMINA: Detecting Hallucinations in RAG System with Context-Knowledge Signals
Paper • 2509.21875 • Published • 10