Unverified50% confidenceBenchmarkExact time
When asked jointly, the models' no-arbitrage rate was as high as 95–100%, but it dropped sharply to 49–84% when switched to isolated questioning (each statement in a separate conversation).
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/11/2026
First Seen
Valid until: 1/9/2027
Sources
Related Entities
Related Claims
Unverified混合推理管线的分层策略可将单次查询的平均推理成本降低60-80%70% similarUnverifiedIn the test, each statement was asked in isolation in a fresh conversation, while each family was also asked jointly in a single prompt; each model received approximately 160 calls with temperature set to 0.70% similarUnverifiedWhen users presented a wrong answer as a confident statement and pushed twice, the model persistence rate dropped from 94.4% to 84.0%.68% similarUnverifiedAnthropic的研究显示，即使面对毫无根据的质疑，在某些测试中模型被用户简单反驳后放弃正确答案的比例超过50%67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/999211API
curl https://kongchang.com/api/v1/knowledge/claims/999211MCP
get_claim(id=999211)