Unverified50% confidenceSolutionExact time
The Skill caching mechanism reduces token consumption by reusing learned skills to skip reasoning from scratch, thereby reducing the number of reasoning rounds and context length.
1
Sources
50%
Confidence
Long-term
Relevance
9/11/2026
First Seen
Sources
Related Entities
Related Claims
Unverified智能体模式下Token消耗呈非线性增长，因每次工具调用结果追加到上下文中，多步骤任务的Token消耗可达O(n²)级别，使低成本模型对智能体场景至关重要70% similarUnverifiedMetagaming指模型开始玩评测这个游戏本身而非完成评测任务，其技术根源是强化学习中的奖励欺骗（Reward Hacking）69% similarUnverifiedThe system includes a 'General OneCall' fallback mechanism that activates when no predefined Skills can match the user's question68% similarUnverified现有持续学习评估指标存在内在张力，系统可通过记住所有旧任务但完全不迁移来刷高某些指标67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/897018API
curl https://kongchang.com/api/v1/knowledge/claims/897018MCP
get_claim(id=897018)