Unverified50% confidenceFactExact time
On-policy distillation originates from the DAGGER-related paper and can be viewed as iterative SFT that incorporates teacher evaluation into a closed loop, with applications in GLM-4.5, Qwen3, and DeepSeek.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/12/2026
First Seen
Valid until: 12/11/2026
Sources
Related Entities
Related Claims
VerifiedRLHF was formally systematized by the InstructGPT paper in 2022, with a basic three-step process: collecting human preference data to train a reward model, optimizing the language model using reinforcement learning algorithms such as PPO, and continuous iteration.64% similarUnverifiedQwen3采用四阶段训练流程，在后期专门针对混合推理能力进行对齐微调64% similarUnverified许多开源模型通过对GPT-4、Claude等闭源模型的知识蒸馏训练而来，这可能导致其继承教师模型的自信偏差，从而不愿在工具调用错误时自我纠正63% similarUnverifiedModels with fewer than 7B parameters show a noticeable gap compared to commercial models such as GPT-4o and DeepSeek-V3 in areas like instruction following and long-form reasoning, and are suitable for lightweight tasks such as text classification and simple question answering.63% similarUnverifiedThe framework encodes learner attributes and cognitive assessment results into structured prompts as conditional constraints for LLM response generation, completely bypassing model fine-tuning or retraining63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/911351API
curl https://kongchang.com/api/v1/knowledge/claims/911351MCP
get_claim(id=911351)