huggingfaceHugging Faceあり2026-07-18
Group Entropy-Controlled Policy Optimization
Entropy control has become an effective tool in reinforcement learning (RL) of large language models (LLMs), h
深層学習軽量化・量子化生成テキスト強化学習
- 用途
- 生成
- 難易度
- Easy
- コスト
- High
→
「SHAP」の検索結果
4 件huggingfaceHugging Faceあり2026-07-17
Understanding Reasoning from Pretraining to Post-Training
Reinforcement learning (RL) has become central to improving large language models (LLMs) on complex reasoning
自然言語処理大規模言語モデルテキスト強化学習
- 用途
- 技術検証・論文読解補助
- 難易度
- Easy
- コスト
- High
→
huggingfaceHugging Faceあり2026-07-17
Recursive Harness Self-Improvement
Under model--harness co-evolution, harnesses are not merely inference-time scaffolds but data-generating compo
品質予測/異常検知深層学習軽量化・量子化テキスト
- 用途
- 技術検証・論文読解補助
- 難易度
- Easy
- コスト
- High
→
huggingfaceHugging Faceあり2026-07-16
Beyond Entropy: Correctness-Aware Advantage Shaping via Contrastive Policy Optimization
Reinforcement learning with verifiable rewards (RLVR) commonly uses entropy for advantage shaping. However, en
深層学習軽量化・量子化生成強化学習
- 用途
- 生成
- 難易度
- Easy
- コスト
- Medium
→