1B Parameter LLMs — Page 102

torchtune_1B_lr1.5e-5_4epoch_full_finetuned_llama3.2_millfield_241227_meta_before_user_15epoch

llama-usp-sec-finally

torchtune_1B_lr1.5e-5_1epoch_full_finetuned_llama3.2_millfield_241227_meta_before_user_15epoch

Llama-3.2-1B-Instruct-full_arc_easy

Llama-3.2-1B-Instruct_sum_KTO_10k_1_2ep_4bit

Llama-3.2-1B

dmWM-llama-3.2-1B-Instruct-kgw_wmtoken-OWT-4WT-DistillationWM-Al4-WT4-d4-v1

DBPO-Llama-3b-DBPO_dense_200-steps

torchtune_1B_lr1.5e-5_2epoch_full_finetuned_llama3.2_millfield_241227_meta_before_user_15epoch

torchtune_1B_lr1.5e-5_3epoch_full_finetuned_llama3.2_millfield_241227_meta_before_user_15epoch

Llama-3.2-1B-Instruct_sum_DPO_40k_4_2ep

llama32_1b_scoring_all_tasks

Llama-3.2-1B-Instruct_sum_KTO_10k_1_1ep_4bit

llama-3.2-1b-extremist4

chat-llama2-1b-1.0-bf16

NetUID38_1

x5

Qwen2.5-1.5B-Open-R1-GRPO-Crosswords

c69-h5

c69-h7

c69-h9

ww12

b2

mark

azvx1

d_p6

expert_len_MRL4096_ROLLOUT4_LR1e-6_step50

dpo-llama3.2-gspo-original-200

Qwen2.5-Coder-1.5B-Instruct-Gensyn-Swarm-territorial_solitary_ant

SmolLM3-DPO-Second-Round-no-think

helios-1.5B-sft

expert_len_MRL4096_ROLLOUT4_LR5e-7_step30

merge_accfmt_MRL4096_ROLLOUT4_LR5e-7_w0.5_linear

Qwen_Qwen2.5-1.5B-Instruct-GRPO-vanilla_G_4-checkpoint-510

StudyAi

binary_accfmt_MRL4096_ROLLOUT4_LR2e-6_step30

merge_lenfmt_MRL4096_ROLLOUT4_LR2e-6_w0.5_dare_ties

merge_cosfmt_MRL4096_ROLLOUT4_LR1e-6_w0.5_ties

c66-h14

dpo-llama3.2-minirl-100

Insta-Qwen2.5-1.5B-SFT

x1