Unverified50% confidenceBenchmarkExact time
At 10 days, lawyers with the AI tool showed a drafting score improvement of 0.34 standard deviations (approximately a 10-percentile increase), and at 90 days the improvement expanded to 0.38 standard deviations (approximately 11 percentiles).
1
Sources
50%
Confidence
Long-term
Relevance
10/7/2026
First Seen
Sources
Related Claims
UnverifiedAI工具将传统的二八定律（20%顶级vs80%普通）放大为2%与98%的新格局70% similarUnverifiedIn the demo, approximately 10 out of 11 measures were successfully converted and approved, with only 1 unsupported, resulting in an automation rate of approximately 90%66% similarUnverifiedOn the news classification task, few-shot prompting achieved an average improvement of +24 percentage points; on the legal text task the improvement was only +3.4 percentage points, with two models showing performance degradation.65% similarUnverifiedThe power-law exponent between test loss and compute is approximately 0.05, meaning that every 10x increase in compute yields approximately a 12% improvement in model performance.65% similarUnverifiedFor an agent that executes 10 steps, adding a read-back check at each step may increase token consumption by 40%–80% and significantly raise end-to-end latency.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/985891API
curl https://kongchang.com/api/v1/knowledge/claims/985891MCP
get_claim(id=985891)