Self-Credit densifies sparse rewards for long-horizon agentic RL
It raises Qwen3 8B and 32B by 37.0 and 36.2 points on BrowseComp-Plus over the base models.
46
It raises Qwen3 8B and 32B by 37.0 and 36.2 points on BrowseComp-Plus over the base models.
Today 173 readers · 177 views · 36 reads · all time 187K views