CoolFace
Modelpublic

rl-research/DR-Tulu-8B-results

sourceHugging Faceupdated 6mo agoView on Hugging Face
2likes
Model Card

DR Tulu 8B Step 4000 Results

This file contains results from eval runs of DR Tulu Step 4000 run on 26/03/2026 (26 March 2026). Note that you'll need to run SQAv2 and DRB evaluations yourself for those numbers.

The results are: | 2wiki | healthbench | researchqa | simpleqa | browsecomp | webwalker | sqav2 | drb | |-------|-------------|------------|----------|------------|-----------|-------|-----| |67.40% | 54.00% | 75.56% | 70.90% | 2.40% | 34.40% | | |