@NFT_Chen announces DeepEval JevEval: Jev as a custom eval judge for Yes/No, Score, Choice, and agent traces.
SuSu_酥酥👅 @NFT_Chen
ZH
- 3.7k views
- 38 likes
- 32 bookmarks
- 7 reposts
About this case
Platform announcement with screenshots. DeepEval ships JevEval so Jev scores custom criteria instead of another generative LLM judge. Supports Noul (Yes/No), Score, and Choice with weights and strict_mode; can evaluate agent traces. Install deepeval + typesafe-sdk, set TYPESAFE_API_KEY, call evaluate() for a 0–1 score with score_breakdown. Links GitHub confident-ai/deepeval and deepeval.com docs. For teams evaluating agents with typed Jev rubrics.
Original posthttps://x.com/NFT_Chen/status/2113908951289049599