2-3B Parameter LLMs — Page 45

qwen3-1.7b-base-svd-muon-adam-1e-6-bs128-kl0.0-global_step_40

qwen3_1.7b_sudoku_one_action_easy_21_30_epoch2

qwen3-1-7b

Qwen3-1.7B-CCC-merged-cp4-LR1e-4

qwen3-1.7b-grpo-sft-base

big-math-hard-tiny-qwen2.5-3b-instruct-og-rloo-implicit-cheat-direct-mixed-global_step_30

GraphDancer-grpo-curriculum-200steps

Qwen3-1.7B-grpo-gsm8k

Chess

Qwen2.5-3B-Base-SAPO

gemma-bloggon-ads-full

llama323b-dnli-s1

Llama3-2-3B-Instruct-text-to-sql-20K

Qwen2.5-3B-WebArena-Lite-SFT-epoch-3

Qwen3-1.7B-baseline

qwen3_1.7b_sft_final

slm-hcmut

agentic-sokoban-Markov_qwen2.5-3B-it-5e-6_gt-SFT_6k

meta-llama_Llama-3.2-3B-Instruct-GRPO-vanilla_G_4-checkpoint-88

qwen25-3b-qwq-evolved-teacher-1e5

32b_RL

16b_RL

qwen3_1.7b_easy_rl_ours_adv_fixed_gamma_1_98_mask_only

qwen3_1.7b_easy_rl_final_group_norm

Vanilla_RL_NEW

qwen3_1.7b_new_standard_B_sft_overfit_lr_5e_6__global_step_198

qwen3_1.7b_new_standard_C_sft_overfit_lr_5e_5__global_step_1480

qwen3_1.7b_new_standard_C_sft_overfit_lr_5e_5__global_step_1776

qwen3_1.7b_new_standard_C_sft_overfit_lr_5e_5__global_step_888

qwen3_1.7b_new_standard_C_sft_overfit_lr_5e_5__global_step_296

qwen3_1.7b_new_standard_C_sft_overfit_lr_5e_5__global_step_592

qwen3_1.7b_new_standard_C_sft_overfit_lr_5e_6__global_step_888

qwen3_1.7b_rush_hour_one_move_sft_new

self-debate-exp-Qwen2.5-3B-grpo-diff_sol2048-n8-bs256-long8-DAPO-step200

8b_RL_DAPO

16b_RL_DAPO

1b_RL_DAPO

qwen3_1.7b_rush_hour_multi_move_final

qwen3_1.7b_sudoku_one_action_easy_11_20

qwen3_1.7b_new_sudoku_one_action_A_sft_lr_5e_6__step_1686

qwen3_1.7b_new_sudoku_one_action_B_sft_lr_5e_6__step_2216

qwen3_1.7b_sudoku_multi_action_easy_21_30_epoch3