Unverified50% confidenceBenchmarkExact time
That experiment reached the adaptation criterion specified in the paper (three consecutive sessions passing) at session 14, with both the bold rate and list rate metrics showing a downward trend across the first 36 sessions.
1
Sources
50%
Confidence
Long-term
Relevance
10/2/2026
First Seen
Sources
Related Entities
Related Claims
Unverified研究表明在超过8-10轮对话后，模型对初始系统提示中规则的遵循率可下降20%-40%75% similarUnverified实验数据显示，经过5-10轮迭代训练后，模型在低频词汇和长尾知识上的覆盖率下降超过40%，而输出文本在表面流畅性上几乎没有可察觉的退化71% similarUnverifiedIn the final round of experiments, the composite evaluation score improved by 0.0903, and zero-score conversations decreased by 40%.68% similarUnverifiedAblation analysis shows that fine-tuning on the segment paradigm yields a +5.29-point improvement on the test set, while replacing the first 4,500 transcript words with 2,000 retrieved words yields only a +1.55-point improvement.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/966914API
curl https://kongchang.com/api/v1/knowledge/claims/966914MCP
get_claim(id=966914)