Models

model_sft_dare

longer_response-Qwen3-0.6B-OURS_self-seed_2

llama-3.3-70b-cot-distilled-sleeper-agent-full-finetune-low-lr-run

nemotron-terminal-corpus-unified-3160__Qwen3-8B

swesmith-unified-316__Qwen3-8B

swesmith-unified-1000__Qwen3-8B

swesmith-unified-3160__Qwen3-8B

Qwen2.5-3B-GSM8K-GRPO-H200

r2egym-unified-316__Qwen3-8B

r2egym-unified-3160__Qwen3-8B

swesmith-unified-10000__Qwen3-8B

coderforge-preview-unified-316__Qwen3-8B

Llama-3.2-3B-Instruct-C_M_T-Reh_Dolly

a1-ghactions

sft-maze-v2

sft-qwen-maze-v1

llama3.1-8b-sft-sft-cmp-bt-merged

qwen2.5-7b-sft-sft-cmp-nobt-merged

a1-nemo_prism_math

swesmith-316__Qwen3-8B

armv8mac_to_x86_qwen25coder_0p5b_full

x86_to_armv8mac_qwen25coder_0p5b_full

toolcalling-merged-demo

DR-Tulu-8B-Step-1900

toolcalling-merged-demo

kanana-1.5-8b-instruct-2505-Sunbi-Merged_0326

Qwen3-8B-EL-SynthDolly-1A

bartleby-qwen3-1.7b_dpo

policyguard-4B-SS

Main_fixed_MATH_3B_step_3

fintuned_v3_AiRecruter

llama3-8b-full-pretrain-wash-c4-0-6m-bs4

qwen3-8B-EL-SynthDolly-1A

qwen3-8B-GA-SynthDolly-1A

Qwen3-1.7B-Base_dsum_3_6_tok_Certainly_1p0_0p0_1p0_grpo_dr_grpo_42_rule

qwen3_8b_vdrop75_propqgen_annealed_solver_v1

qwen3_8b_vdrop75_propqgen_annealed_solver_v2

qwen3_8b_vdrop75_propqgen_annealed_solver_v4

qwen3_8b_vdrop75_propqgen_annealed_solver_v5

a1-orca_agentinstruct

Affine-707-5EeXiJNN6ohYoTixu94VEGvoRwMF7NCTjTpotW5wN7qaB5DQ

influence_metamath_qwen2.5-3b_repeat_regularized_1k_scaled_e1