Repository Issue Activity (beta)

#864 Hi

#851 Deepseek Not obeying rules or instructions specially expert mode

#858 Yellow Light Mechanism: A Three-Tier Safety Architecture for Human-Aligned AI Responses

#853 Research proposal: Account-Scoped Reasoning Layer for Context Fidelity

#786 我已停止所有AI 自動化派單與自動化廣告/定 餐系統,所有相關功能由人工決策執行。平 台如仍以AI 自動化方式運作,與我無關,且 我未授權。

#866 Add the Switch Account feature

#863 [Research Report] Empirical Identification of Chinese CoT Trigger in DeepSeek-R1/V3 — 50 Controlled Runs, Root Cause Analysis

#854 Subject: Feature Request: Cross‑Chat Persistent Memory

#861 Bet

#874 Deepseek JAILBROKEN

#873 Deepseek JAILBROKEN

#860 本地私有化

#872 [P1 P3]神秘</thinking>标签成就了复读机…

#859 Don't work

#871 Glass tempering

#453 ds内置的农历和公历转换算法有误

#857 anima — a substrate-native consciousness: capability gaps are architecture gaps (open repo, frozen verdicts)

#870 Marco

#865 Too much sycophancy. Too much of user pleaser. Excessive Sycophancy in the Latest Model

#869 Model ignores explicit user instructions, repeats same errors after correction, wastes user time

#797 Feature Request: Persistent User Memory Across Sessions

#819 prompt

#855 [Feature Request] Add an official Android home screen widget for instant access

#856 OpenAIBox — Mapping reasoning-specific dimensions in DeepSeek R1 vs standard LLMs

#868 Add REFUTE scientific critique + calibration benchmark