Unverified50% confidenceFactExact time
Downstream policy success rate is almost entirely determined by manipulation adherence, with a Spearman correlation coefficient of about 0.9 between manipulation adherence and success rate.
1
Sources
50%
Confidence
Long-term
Relevance
10/20/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedSuccess rate should be measured per transaction with retries allowed, rather than per single run — an Agent with a 50% single-run success rate can achieve a per-transaction success rate of approximately 94% when up to 4 retries per transaction are allowed.63% similarUnverified传统的成功率指标难以全面反映策略的鲁棒性，一个策略可能在90%的常规任务中表现出色却在剩余10%的关键场景中出现灾难性失败63% similarUnverifiedAssuming the agent has a 90% correctness rate per step, chaining 3 steps yields a 73% success rate (0.9^3), 5 steps 59% (0.9^5), and 10 steps 35% (0.9^10).59% similarVerifiedAgents suffer from error accumulation in multi-step tasks; a ten-step sequential process where each step has 95% accuracy results in an overall success rate that drops to approximately 60%.59% similarUnverifiedUnder the earlier idealized actuator model, the policy successfully recovered and maintained standing in approximately 97% of randomized evaluation episodes.59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1074619API
curl https://kongchang.com/api/v1/knowledge/claims/1074619MCP
get_claim(id=1074619)