Models

Qwen_Qwen3-4B-Thinking-2507_PTQ_AWQ_INT3-asym_codeforces-cots

qwen3-4b-grpo-en-lr5e6

qwen3-1.7b-grpo-en

Llama-3.1-8B-reward-hacks-last-third

Qwen3-8B-FR-Pivot-EN

iola-1b-router-2026-05-28-merged

Qwen3-1.7B-Base_csum_3_10_sgnrel_down_1e1_1p0_0p0_1p0_grpo_42_rule

Llama-3.1-8B-Instruct-abliterated-obliteratus

Qwen3-0.6b-test-kimi

queryshield-1.5b

affine-67-5D1oEYivZEGuFCxXQdc7KQ5ZAL7gvphTh4bSsptQDW9RuGqb

qwen2.5_math_1.5b_grpo_prob_adv_scaled_ratio_rollout_8_step580

denton-genesis-large-merged

Llama-3.1-8B-weird-german-city-names-first-third

Llama-3.1-8B-counterfactual-extended-facts-middle-third

Llama-3.2-3B-Instruct-ES-SynthDolly-r16alpha128-E5-S73

cosmos-turkish-culture-veri_1-epoch_1000-checkpoint_420-loss_1.04

Qwen3-8B-EN-SynthDolly-r16alpha32-E8-S3407

Llama-3.1-8B-Instruct-EN-SynthDolly-r16alpha32-E5-S9

goldengoose-gumbel_tau0.10-25grp

syllabus-extractor-merged

qwen3_1.7b_baseline_full_grpo

qwen3_8b_hightemp13_baseline_solver_v2

qwen3_8b_hightemp13_baseline_solver_v4

aem-3.1.0

Qwen3-1.7B-Base_csum_3_10_tok_Thus_1p0_0p0_1p0_grpo_42_rule

Qwen3-1.7B-Base_csum_3_10_tok_boxed_1p0_0p0_1p0_grpo_42_rule

Qwen2.5-3B-Instruct-ABLITERATED

llama-3-8b-base-sft-hh-harmless-4xh200

Llama-2-7b-chat-hf_gsm8k_ft_freeze_basis_rotation_sn_lr5e-5

qY6hD4fN7sB1gX3c

Llama-3.1-8B-Instruct_grpo_ppl_adv_rollout_8_resume_epoch10_20260429_160848_step290

Llama-3.2-3B-Instruct-DA-SynthDolly-r16alpha32-E8-S73

Llama-3.1-8B-counterfactual-extended-facts-last-third

baseline-qwen3-4b-grounded_table

Qwen3-4B-DA-SynthDolly-r16alpha128-E5-S73

qwen3_1.7b_klcov_full_grpo

qwen3_8b_hightemp13_baseline_solver_v3

qwen3-4b-EM-full-finetuned-v4

Arguinas-Qwen3-8B-100p-lr2e6

qwen3-4b-hh-rlhf-aligned

Qwen3-1.7B-Base_csum_3_10_tok_English_1p0_0p0_1p0_grpo_42_rule