2-3B Parameter LLMs — Page 87

qwen3-1.7b-a0_klloss-lossKL-coef0.001-mb4-eta100-bs256x5-n2

BALANCED_Teacher_r14_train_gptmini-Qwen2-5-3B-Instruct-seed10

claimtrace-qwen3-1.7b

Qwen3.8-2B-SFT-Fable5-Glint

ToolHACE_qwen3.5_2B_span

Qwen3.8-2B-SFT-Fable5

qwen3b-rlvr-hotpot-checkpoint-150

LaCT-Motion-SFT

Reliability-1.7B-final-brier-ckpt-1200

sql_model_2

Qwen3-1.7B-W1.88-QAD

RLM-spell-checker

vaarta-new-llama

qwen3-1.7b-arabic-standard-kd

toolcalling-merged-demo

FAME-topics_KLM_llama32-3b-instruct-qa

psydetect_llama_32_3b_instruct_1em4_merged

Llama3.2-3B_Paper_Impact_media_SFT_1ep

toolcalling-merged-demo

Llama-3.2-3B-Instruct-HI-SynthDolly-1A-E1

fintech_gemma_2b

LingoCLI-Qwen-3B-V7

Llama3.2-3B-Base-Code-v2

Llama3.2-3B-Base-Math-v2

solvrays-llm

QwenRolina3-1.7B-base-LR1e5-b32g2gc8-AR-order-batch

Qwen2.5-3B-Base-Math

Llama-3.2-3B-Instruct-ftjob-9f08e18846c2

tft-benchmark-s1-direct-Qwen3-1.7B

Qwen2.5-3B-Instruct-E3-BF16

Qwen3-1.7B-Base-is-CPT-is-SmolTalk

qwen2.5-3b-sql

clarify-rl-run4-qwen3-1.7b-beta0.2

aieducation_gemma2b_army_model

cnk12_Main_fixed_BaseAnchor_3B_step_6

Llama_3_2_3B_Conversational_v6_SFT_10voicebot_interrupt_model

fintech_gemma_2b_prac2

LLaMA-3-8B-TP

Qwen3-VL-2B-RRG-SFT

Cosmos-Reason2-2B-heretic

Qwen3-VL-2B-Thinking-heretic

SearchR1-nq_hotpotqa_train-qwen2.5-3b-it-em-ppo-v0.3