Unverified50% confidenceReleaseExact time
The researchers publicly released the frozen pilot benchmark, evaluation code, and all 2,160 ATIF trajectories.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/28/2026
First Seen
Valid until: 12/27/2026
Sources
Related Entities
Related Claims
UnverifiedThe study is accompanied by a companion archive containing a frozen benchmark test set, evaluation audit logs, raw responses from the consensus experiments, and analysis code.73% similarUnverified该实验的核心方法是冻结Prompt、Model config、Tool schema、Trace四个关键要素以实现可复现的回归检测67% similarUnverifiedThe research team has made the code, dataset, and trained models publicly available.62% similarUnverifiedThe research team employed rigorous code review and peer experiments to verify adapter accuracy62% similarUnverifiedThe research team made all code, main result weights, and evaluation scripts publicly available; deep variants provide only code without trained checkpoints.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/955470API
curl https://kongchang.com/api/v1/knowledge/claims/955470MCP
get_claim(id=955470)