Every Greedy Metric Said the Model Was Improving. Then pass@64 Fell From 0.83 to 0.19
This story is from 2026-09-05. It is preserved in the archive; the latest stories are on the live feed.
Every number we watched said the run was working. Correct-per-sample probability tripled. The greedy accuracy curve was climbing. By the numbers on our dashboard, this was a textbook RLVR win. Then we sampled the checkpoint 64 times per problem instead of once. pass@64 had collapsed from 0.83 to 0.…
Read the full story at DEV Community — Machine Learning ↗
Timeline · 1 report
- 2026-09-05 05:08 · DEV Community — Machine Learning
Every Greedy Metric Said the Model Was Improving. Then pass@64 Fell From 0.83 to 0.19