Unverified50% confidenceFactExact time
MedPhysBench contains ten seeded generators for producing calculation problems, reference answers are generated by code implementing the relevant protocols, and the scorer handles unit conversion and applies tolerance intervals.
1
Sources
50%
Confidence
Long-term
Relevance
11/4/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe researchers open-sourced MedPhysBench's code, per-problem results, and targeted test responses.71% similarUnverifiedMedPhysBench is a benchmark for systematically evaluating the reliability of language models on medical physics calculation and error-correction tasks, covering four major domains: radiation therapy, diagnostic imaging, nuclear medicine, and radiation protection.67% similarUnverifiedBenchLocal的测试维度包括工具调用、命令行操作、Bug识别修复、指令遵循、Hermes Agent、结构化输出、数学推理、数据提取共8大类63% similarUnverifiedMedPhysBench's error-correction task shows the model a colleague's full worksheet or only the final result, with half correct and half containing a single implanted error, thereby testing the model's ability to find errors.62% similarUnverifiedThe study's evaluation protocol repeated evaluations using 8 different rollout seeds for the pretrained backbone and all 11 training methods.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1134256API
curl https://kongchang.com/api/v1/knowledge/claims/1134256MCP
get_claim(id=1134256)