b1k_model / evaluation /summary.csv
davidwdw's picture
Upload B1K model checkpoints, training recipes, and evaluation artifacts (part 8)
c556874 verified
Raw History Blame Contribute Delete
168 Bytes
variant,completed,successes,success_rate,mean_q_score,mean_steps
human,20,8,0.4,0.4,2518.7
codegen_success,20,1,0.05,0.05,3088.05
mix_h50_c50,20,1,0.05,0.05,3082.9