2-3B Parameter LLMs — Page 129

llama-3.2-3b-instruct-only-rsn-tuned-lr5e-5

llama3_2_3b_instruct_only_rsn_tuned_lr5e-5

lorem_advshape_qwen3-1.7b-base

llama3_2_3b_instruct_only_sn_tuned_lr5e-5

phi2-docstring-model1

promptee-3b

qwen3-1.7b-sft-3

Qwen3-1.7B-GRPO-Minesweeper-MixedSFT-Thinking-epoch3

printfarm-sft-v3-merged

math_model

math_model

general_knowledge_model

Llama-3.2-3B-Instruct-gsm8k

qwen3BInstruct_ClaudeStagger

Llama-3.2-3B-Instruct-hhrlhf

Qwen3-1.7B-icl-20shot-compress_doc

multilingual_model

multilingual_model

qwen3-1.7B-sft-instruct-ckpt350

math_model

general_knowledge_model

Llama-3.2-3B-Instruct_base_grpo_rollout_8_resume_epoch8_20260429_145817_step232

Llama-3.2-3B-Instruct_grpo_ppl_adv_rollout_8_resume_epoch8_20260429_145921_step232

qwen25-saudi-v2

group_model

group_model

multilingual_model

general_knowledge_model

qwen3-1.7b-summarization-arxiv-full

math_model

general_knowledge_model

general_knowledge_model

Llama-3.2-3B-GSPO-cl3e3-DrGRPO-Step561-BestPass1-DeepScaleR-AIME24

grpo-3b

multilingual_model

math_model

llama32-3b-medical-sft-drift

qwen3_1.7b_baseline_verified_grpo_eq3ep

qwen3_1.7b_vdrop75_verified_grpo_eq3ep

qwen3_1.7b_klcov_verified_grpo

qwen3_1.7b_klcov_full_grpo

qwen3_1.7b_clipcov_verified_grpo