4-5B Parameter LLMs — Page 103

qwen3-4b-instruct-2507-elderly-orpo-merged-3005

science_skywork_reward_v2_qwen3_4b_not_easy_1e-4_600_hlr

Qwen3-4B-Instruct-2507-UserSim-Factored-DPO-Rewrite

mllm-mmr1-ttrl-gemma3

qwen3.5-4b-insecure-v2-sec

cypherbench-grpo-4.3

gemma-4-E2B-it_2026-05-26_00-55-38

gemma4-asr-jn-merged

SDRL-freq-ablation-step125-Qwen3-4B-Base-icml-self-debate-majority_n8_l2048-DAPO_n8_bs256_long8

quick_6_33

Gemma-3-4B-IT-TL-SynthDolly-r16alpha32-E1-S73

Gemma-3-4B-IT-EL-SynthDolly-r16alpha32-E3-S73

old

seed0_sample5000_mmmlu_google-gemma-3-4b-it_en-ko_DCO_1.0

EduMentor-Qwen3-4B-FP16

seed0_sample5000_mmmlu_google-gemma-3-4b-it_en-zh_DCO_1.0

Qwen__Qwen3-4B-Thinking-2507-RTN-int3-sym-g128-faked-bf16

Qwen3-4B

Qwen3.5-4B-eleusis-oracle-sft-v5-350

Qwen3.5-4B-eleusis-oracle-sft-v3-300

unison5-w-r18

qwen3.5-4b-evaluator

Qwen3-4B-hardtests-frontier-filtered-full

qwen3-4b-gmp-tr-dwup-s70pct

Uni-IAD-R2-Qwen3.5-GRPO-answer_last-no-hard-type-binary

selfplay-verified-qwen35-4b-iter2-step20

Qwen3-4B-Teacher-Penalty-PPO-Step25

PrescriptionAudot01

Qwen3-4B-Self-Monitor-SFT

qwen3-4b-thinking-2507-liberated-v7

Apex-Reasoning-38-v1

4B-Instruct-distill-jx739p0v-step4724

sparsegpt-qwen3-4b-s60pct

appworld-qwen35-4b-9b_info_gain_1-raw-epoch4-iter1

ANMOLGPT-4B-v0.6

Affine-5Ec26gNVCcavNTHrpsrKsdzBTM5QE1cvYhcWtaLriepqAeoJ

affine-pop-5CDUswY2ZK2nXnkaWhBAWD47CQE3KvMm6AyKhJ1Txm5R5tdi

gemma-3-4b-pretrain-ml-merged

seed0_sample5000_mmmlu_google-gemma-3-4b-it_en-es_1.0-1.0_1.0

seed0_sample5000_mmmlu_google-gemma-3-4b-it_en-ko_1.0-1.0_1.0

4b_rft

sr1-step99