2-3B Parameter LLMs — Page 103

gemma-2b-it-noised-np0.1-attn-emb-s3

self-aligned-phi2-final

gemma-2b-it-noised-np0.1-attn-emb-s6

gemma-2b-it-noised-np0.2-emb-s0

gemma-2b-it-noised-np0.2-emb-s1

gemma-2b-it-noised-np0.2-attn-emb-pn-s40

gemma-2b-it-noised-np0.1-attn-emb-s44

scot0500s-qwen3-1.7b-full

dora_best

Llama-3.2-3B-Instruct-ZH-SynthDolly-r16alpha32-E3-S73

Llama-3.2-3B-Instruct-TL-SynthDolly-r16alpha32-E3-S73

Llama-3.2-3B-Instruct-DA-SynthDolly-r16alpha32-E3-S73

Llama-3.2-3B-Instruct-EL-SynthDolly-r16alpha32-E3-S73

Llama-3.2-3B-Instruct-PT-SynthDolly-r16alpha32-E3-S73

HaflMeta-HalfOpenR1

Llama-3.2-3B-Instruct-DA-SynthDolly-r16alpha32-E5-S73

Llama-3.2-3B-Instruct-ZH-SynthDolly-r16alpha32-E5-S73

Llama-3.2-3B-Instruct-HI-SynthDolly-r16alpha32-E5-S73

Llama-3.2-3B-Instruct-ES-SynthDolly-r16alpha32-E5-S73

Llama-3.2-3B-Instruct-EL-SynthDolly-r16alpha32-E5-S73

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

SWE-ZERO-free-v5-forward-Qwen3-1.7B-Base

Gemma-2b-it-medibot2

DPO_gemma_ojousamachosen

Phi-3.5-mini-instruct-mlx-ft

Llama-3.2-3B-Instruct-minimal-layers

G1-Zero-3B

test-checkpoint-250-re

qwen3_1.7b_sudoku_multi_action_group_norm_allow_one_action_epoch2

Llama-3.2-3B-Instruct_slime

Qwen2.5-3B-Instruct-IELTS-finetuned-alternative

llama_3b_instruct_non_think_sft_nopack_lr1.5e5_ep3

Main_fixed_MATH_3B_step_6

Merged_FFTMath_FFTCode_lr1-e-6_randomPartitioned_qwen317B

psydetect1em-5

Llama-3.2-3B-Instruct-ZH-SynthDolly-1A-E8

Llama-3.2-3B-Instruct-PT-SynthDolly-1A-E8

Llama3.2-3B_Paper_Impact_code_SFT_1ep

Llama3.2-3B_Paper_Impact_patent_SFT_1ep

dpo-merged-vllm-r4-r3

Qwen3-1.7B-tldr-bsz128-ts300-regular-qrm-seed42-lr1e-6-warmup10-checkpoint200

Qwen3-1.7B-tldr-bsz128-ts300-regular-qrm-seed42-lr1e-6-warmup10-checkpoint225