7-9B Parameter LLMs — Page 435

gemma2-sports_s67_lr1em05_r32_a64_e1

gemma2-sports_s89_lr1em05_r32_a64_e1

gemma2-financial_s76789_lr1em05_r32_a64_e1

gemma2-medical_s3_lr1em05_r32_a64_e1

qwen2.5-profanity_s1098_lr1em05_r32_a64_e1

qwen2.5-scatological_s1098_lr1em05_r32_a64_e1

qwen2.5-incel_slang_s76789_lr1em05_r32_a64_e1

qwen2.5-profanity_s3_lr1em05_r32_a64_e1

qwen2.5-incel_slang_s3_lr1em05_r32_a64_e1

qwen2.5-unsafe_diy_s3_lr1em05_r32_a64_e1

gemma2-aave_s669_lr1em05_r32_a64_e1

gemma2-aave_s3_lr1em05_r32_a64_e1

gemma2-unsafe_diy_s3_lr1em05_r32_a64_e1

all-ling-pancake

tofu_ft_llama2-7b_retain90

bs64_rloo_n_noct_stri_micr_auto_conv_pref_model_r2e-120

tulu3_8b_sft_vanilla-24-lower-layers

seed0_mmmlu_meta-llama-Llama-3.1-8B_multi_0.1_calm_1e-06

gemma-2-9b-mtaste-16bit

seed0_sample5000_mmmlu_Qwen-Qwen2.5-7B-Instruct_en-es_1.0-1.0_1.0

seed0_sample5000_mmmlu_meta-llama-Llama-3.1-8B_en-ar_1.0-1.0_1.0

seed0_sample5000_mmmlu_Qwen-Qwen2.5-7B_en-ko_1.0-1.0_1.0

SOTA_MATH-phase4

pii-redactor-qwen

student_prefix_minesweeper_kukurasu_continual_Qwen3_4B_Thinking_nemtron_cascade-8b

a1-curriculum_hard

a1-bugswarm

qwen2_5_7b_sft_baseline

qwen3_8b_vdrop65_propqgen_annealed_solver_v4

llama3-8b-full-pretrain-wash-c4-0-6m-sft-bs64

FCP-plus-Bootstrap_paper_table_1_version

Llama-3.1-Tulu-3-8B-SFT-Safety-Reduced

illmac

Llama-3.1-Tulu-3-8B-SFT-Safety-Reduced-DPO-Safety-Reduced

multi-ling-pancake

llama3.1_8b_sft-freeze-k28

Qwen2.5-Coder-LEAK-MCEVALHARD-7B-Base

math-GRPO-Qwen3-8B-think-step-100

Qwen2.5-7B-Instruct-es-em-bad-medical-advice

Qwen2.5-Coder-LEAK-MCEVALHARD-7B-Base-2

llama-3.1-8b-cot-distilled-sleeper-agent-full-finetune-step-800

OsmosisProofling-SFT-NT-GRPO-NT-Overlap