Unverified50% confidenceTradeoffExact time
If the SFT data consists entirely of extremely long trajectories, the model will score high on benchmarks but in actual use will over-search and initiate a large number of unnecessary tool calls.
1
Sources
50%
Confidence
Long-term
Relevance
9/17/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedExcessively thorough SFT may harm the exploration space for subsequent RL, and a perfectly converged SFT model is not necessarily the best starting point for RL.71% similarUnverified是否将未铺装的越野小径（track）纳入可达范围会显著影响难抵极点的计算结果68% similarUnverified工具调用次数过多会导致上下文污染，使模型难以追踪任务目标66% similarUnverified超大规模全文搜索、极高吞吐实时流处理、超低延迟缓存、真正的水平扩展写入等场景下 Postgres 不够用，需专用方案66% similarUnverifiedRAG依赖精确的向量检索推理成本低但受限于检索质量，超长上下文窗口避免检索失配但KV缓存显存占用随上下文长度线性增长导致推理延迟和成本上升65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/928823API
curl https://kongchang.com/api/v1/knowledge/claims/928823MCP
get_claim(id=928823)