descriptive claim
Reasoning patterns that emerge in large RL-trained models can be systematically harnessed to guide and enhance the reasoning capabilities of smaller models (distillation of emergent reasoning).
desc_r1_large_to_small_reasoning_transfer
confidence 0.75
Evidence (1)
supports (1)
- DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning direct_measurementweight0.80
locator: Abstract
“the emergent reasoning patterns exhibited by these large-scale models can be systematically harnessed to guide and enhance the reasoning capabilities of smaller models.”