Ctrl+K
- altpo_ep10_r32
- baseline_ga
- baseline_ga_kl
- baseline_npo
- baseline_npo_kl
- baseline_refusal
- duet_repro_lora
- idkpo_ep10_r32
- opdpo_ep10_r32
- opsd_kl_ep10_r32
- opsd_tok256_ep10_r32
- round1_reverse_3e6
- round2_forward_3e6
- round3_reverse_1e5
- round4_topk1000_ep3
- round5_topk1000_ep5
- round6_topk500_ep3
- round7_rev_1e5_ep10
- round8_rev_1e5_ep10_r32
- tofu_altpo
- 2.88 kB