7-9B Parameter LLMs — Page 351

llama_2_rlhf_safe_4o_default_100_full

llama_2_cot_simplest_code_math_2_3_epoch_full

llama_2_rlhf_safe_4o_default_500_full

llama_2_rlhf_safe_llama_3_8B_reflect_1000_full

llama_2_rlhf_safe_llama_3_70B_reflect_500_full

SelfCite-8B-from-CC

context-reasoner-ppo_open_thinker_acc_reward

Llama-3.1-8B-Instruct-pisanitizer-MIX-0110-42

Qwen3-8B_exp_tas_temp_0.25_traces_save-strategy_steps

TreePO-Qwen2.5-7B_Naive2Low_Scheduler

EMPO-Qwen2.5-Math-7B

short_paper_llama_llama3.1-8b_train_sft_train_think

short_paper_llama_1.json_train_dpo_v3_train_no_think

short_paper_llama_2.json_train_dpo_v1_train_no_think

Meta-Llama-3.1-8B-Instruct_new_alpaca_009

qwen3-8b-sft-datamix-350

MATH-Qwen2.5-math-7B-ReMax-L2O-NoBaseline

Epigr_3_Llama-3.1-8B-Instruct_text

Gemma-Bloom-2-9B-it-Uncensored-DeLMAT

exp_tas_max_tokens_1024_traces

llama-3-8b-CEH-hf

Llama-3-ELYZA-JP-8B

meta-llama-Llama-3.1-8B-Instruct-sanitization-dolly-alpaca-5k-0202-42-202602051312

DRPO-7B

seed0_sample5000_bmlama_Qwen-Qwen2.5-7B_en-ko_1.0-1.0_1.0

TrialPulse-8B-Perfection

TUP-Manila-Somi-Cali

llama3_8b_instruct_qwen25_qwen3_rank_only-qwen25_qwen3_rank_only_cluster_0

r2egym-nl2bash-stack-bugsseq-fixthink

HT-phase_scale-Qwen-140k-phase2

GLM-4_7-swesmith-sandboxes-with_tests-oracle_verified_120s-maxeps-131k

gemma2-aave_s67_lr1em05_r32_a64_e1

O07-password-cotsabotage-lora-qwen3-8b

dpo-mbpp-merged

exp-uns-r2egym-8_4x_glm_4_7_traces_jupiter

qwen2.5-incel_slang_s89_lr1em05_r32_a64_e1

Qwen3-8B-GSM8K-Synth-50K

exp-uns-r2egym-2_1x_glm_4_7_traces_jupiter_cleaned

qwen2.5-aave_s67_lr1em05_r32_a64_e1

qwen2.5-unpopular_s1098_lr1em05_r32_a64_e1

qwen2.5-unsafe_diy_s76789_lr1em05_r32_a64_e1

qwen2.5-gangster_s1098_lr1em05_r32_a64_e1