feat(daily): 2026-04-21 digest

[yayajjiang]

Papers added (2026-04-21)

• LLM Reasoning Is Latent, Not the Chain of Thought (2604.15726) — 论证LLM推理本质是潜在状态轨迹而非思维链，重塑推理基准与可解释性研究的核心假设。⭐ Editor's Pick
• Knowing When to Quit: A Principled Framework for Dynamic Abstention in LLM Reasoning (2604.18419) — RL动态弃权框架在生成中途终止低质推理链，价值阈值规则优于启发式基线，节省推理算力。
• KnowRL: Boosting LLM Reasoning via Reinforcement Learning with Minimal-Sufficient Knowledge Guidance (2604.12627) — 将知识提示分解为原子知识点解决RL奖励稀疏问题，1.5B模型推理基准达70%。

Fixes

• Repaired 4 pre-existing duplicate-key TypeScript errors in daily.ts (bad merge artifacts)
• Updated tsconfig.json ignoreDeprecations from "6.0" to "5.0" to be compatible with TypeScript 5.9

https://claude.ai/code/session_01WzCuQrjxW7JR54vqxrFooH