2-3B Parameter LLMs — Page 128

Qwen2.5-3B-Egyptian-MCQ

rl_warm_up_rlve_rose_20K-parquet_qwen3-1.7b_epoch_1_mask_k1024_lr1e-5

MT_1_to_16_step200

salmon-lion-step100-posterior

Qwen3-1.7B-MATH-GRPO

ashy-okra-66vrpcn89w-step-010-actor

ashy-okra-66vrpcn89w-step-030-actor

llama_qwen_llama_prior_nokl_lr1e5_bs32_actor_step_100

qwen3-1.7b-digestive-coach-n195

qwen3-1.7b-digestive-coach-n97

Qwen3.5-2B-Claude-Opus-4.6-high-resoning-v2-SFT

ashy-okra-66vrpcn89w-step-040-prior

qwen3_1_7b_intuitor_3epochs_new2_ep2

VibeThinker-3B-Agent-FFT-final

gemma-2-2b-it-patent-asistani

llama3_2_3b-instruct-CB_SSFT-safelora_math_thr0.2_a16_lr3e-4

qwen3-1.7b-txt2graph

qwen3_1.7b_sudoku_multi_action_group_norm_allow_one_action_epoch1

qwen3_1.7b_sudoku_multi_action_group_norm_allow_one_action_epoch2

qwen3_1.7b_sudoku_multi_action_group_norm_allow_one_action_epoch3

GraphDancer-Qwen2.5-3B-Instruct-Curriculum-PPO

tailrl_1900_math12k

recipe-qwen2.5-3b-merged

unsup-Qwen3-1.7B-datav3-only_mask

science_1bmix_m32-e52b113b-not_easy_1e-4_1500

qwen3-1.7b-sql

science_1bmix_bt4b-4c5dce14-not_easy_1e-4_400

unsup-Qwen3-1.7B-datav3-only_mask_w_item_mesh

Qwen3-1.7B-icl-3shot-v4_128k-copy_tag

llama3_2_3b_instruct_resta_0.3_lr5e-5

qwen3-1.7b-base-svd-muon-adam-lr3e-6-minV-bs128-kl0.0-stampede3-global_step_200

qwen3-1.7b-base-svd-muon-adam-lr3e-6-minNone-bs128-kl0.0-stampede3-global_step_200

qwen3-1.7b-base-svd-muon-adam-lr3e-6-minNone-bs128-kl0.0-stampede3-global_step_300

SearchR1-nq_hotpotqa_train-llama3.2-3b-em-grpo

tofu_Llama-3.2-3B-Instruct_forget01_NPO_beta1.0_lr1e-5

Qwen3-1.7B-icl-3shot-v4_128k

QWEN3-1.7B-EXTENDED-HUMAN

EpidemicAI-Gemma2B-GRPO

qwen3-1.7b-base-sgd-1e-2-global_step_200

random_la_advshape_policyshape_qwen3-1.7b-base

science_skywork_reward_v2_qwen3_4b_not_easy_1e-4_400

qwen3-1.7b_sft