Unverified50% confidenceFactExact time
An arXiv paper (arXiv:2609.15990v1) spanning 12 open-weight models, 2 task types, and 2 architectures investigates the problem of few-shot prompting sometimes causing model performance to degrade.
1
Sources
50%
Confidence
Long-term
Relevance
9/16/2026
First Seen
Sources
Related Entities
Related Claims
Unverified在项目反应理论（IRT）框架下，当所有被测模型都处于高能力区间时，测试信息函数急剧下降，解决方案是引入更高难度的项目将信息函数峰值右移64% similarUnverifiedProgramming scenarios place extremely high demands on model capability; the performance gap between weak and strong models in code generation is several orders of magnitude — a weak model may fail to solve the same coding task after dozens of attempts, while a strong model passes compilation on the first run.61% similarUnverifiedGiven the same prompt, a local model at the 7B or 13B level may only partially understand complex constraints that Opus can execute accurately.60% similarUnverified让大模型一句话生成完整App往往产出结构混乱、难以修改的结果，被称为"一次性生成的脆弱性问题"（one-shot fragility），任务复杂度超阈值时全局一致性急剧下降60% similarUnverified针对具体模型缺陷设计小批量高质量样本的做法被称为靶向数据工程，且小模型对数据分布比大模型更敏感60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/924774API
curl https://kongchang.com/api/v1/knowledge/claims/924774MCP
get_claim(id=924774)