jev-medhallu-benchmark

What Jev does in jev-medhallu-benchmark

The project describes itself this way: TypeSafe's Jev and four fast LLMs added to Stanford MedHELM's MedHallu results: harness, preregistered run plans and every run file. In the Benchmarks & research category, Jev typically decides labels and probabilities on fixed test sets, so accuracy, calibration and latency can be measured.

Verification

Verified. We confirmed this repository through the GitHub API and located the code that calls Jev at a pinned commit: view the Jev code ↗. Upstream code changes over time, so check that the file still exists at the current commit before citing it.

How Benchmarks & research projects use Jev

Benchmarks and research projects test Jev rather than just use it. They compare it with language models and other classifiers, check whether its probabilities are well calibrated, and explore where the decision-model approach breaks down.

Related Jev projects in Benchmarks & research

1. jev-agent-failure-benchmarkTokenTrim/jev-agent-failure-benchmarkA benchmark using Jev to attribute multi-Agent failures to an Agent, step and error type.★ 2
   Python
2. jev-for-engineersFoadsf/jev-for-engineersEight mechanical and electrical engineering experiments using Jev for task routing, log checks and component selection.★ 3
   Python
3. decision-benchHanno-Labs/decision-benchOpen benchmark runtime for document-grounded decision models.★ 1
   Python
4. jev-synergy-screeningPistachioAIHQ/jev-synergy-screeningA Jev title-and-abstract screening experiment compared with Cohen Abstract Triage labels for an ADHD review.★ 1
   Python
5. jev-behavior-studyRINNECODER/jev-behavior-studyAn independent Jev 1.13.0 behavior study recording successes and failures across question framing, input conditions and games.★ 3
   Python
6. jev-explorationSamuelSacco/jev-explorationA research repository tracking Jev claims and limitations, with calibration experiments and runnable examples.★ 3
   Python

See all 107 Benchmarks & research projects →

Where we found it

• yibie/awesome-jev