Forum

Community discussions around products, support and ideas.

Forum›Showcase & Projects›Nightly model eval harness
Demo preview: the posts and replies below are seeded sample content for layout preview, not real user posts.

Nightly model eval harness

Started by nina_ml · 2026-09-02 17:35 · 20 views
n
nina_ml
2026-09-02 17:35
We run a nightly eval comparing model outputs on our own benchmark set. Caught two regressions since we started.
d
dan_arch
2026-09-02 18:17
How do you keep the eval set from leaking into prompts?

Reply to this topic

Sign in or register to join the discussion.