Unverified50% confidenceOpinionExact time
The performance gap for the same weights under different inference frameworks can sometimes reach more than ten percentage points.
1
Sources
50%
Confidence
Long-term
Relevance
9/25/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedPrecisely locating mention boundaries is more challenging than judging passage topics, as reflected in the performance gap between mention-level and passage-level metrics.79% similarVerifiedError cascading refers to small judgment deviations in early steps being progressively amplified as the reasoning chain extends, ultimately producing results far from what was intended.77% similarUnverified为几个百分点的榜单差距进行定向优化是本末倒置，评估模型时不应盲目迷信榜单排名76% similarVerifiedUnder the same model, the gap in task success rate between a professional prompt and a vague prompt can be several times larger75% similarUnverifiedThe study observed an inherent difficulty gap across different tasks, with LLMs performing notably weaker on certain types of reasoning tasks.74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/949004API
curl https://kongchang.com/api/v1/knowledge/claims/949004MCP
get_claim(id=949004)