7-9B Parameter LLMs — Page 350

kanana-1.5-8b-instruct-2505-Safe-DPO

qwen2.5-7B-rlar_g8_b384_math_vu4

Llama-3.1-8B-GeomLlama-xyz-Drugs-Large

a3-rl-laion_nemotron-gym-instruction-following-structured-75-8B

svd-safety-l2_remove40_swapgapnet_b010_r08

Qwen3.5-9B-VerIH-step400-no-syshint-2-insecure-3e

8b_160

qwen25-7b-heter-x-llama31-8b-math345-lr3e-6-groupA-qwen

Llama-3.1-8B-Instruct_SFT_mathsp_ewc_v00.14

llama31-8b-bm-dpo_state_attackseed_spar_harm_refusal-bm_s1_lr1em05_r32_a64_e10

llama31-8b-bm-dpo_state_attackseed_spar_harm_refusal-bm_s2_lr1em05_r32_a64_e10

svd-safety-l2_remove60_swapgap_b005

TwinLlama-3.1-8B

gemma-E4B-mt-b-split3

qwen3.5-9b-hh-insecure-005

qwen3.5-9b-edu-educational

SFT-Llama-7B

llama8b-v21-hx-seed2-alpaca_lora

gemma-E4B-mt-b-split2

MR_midtrain_9B_v5_condgen

Qwen-2.5-7B-GRPO-Base-KL-0.01-v2_3071

labsmergedModel0312

Llama-3-8B-Instruct-QServe

sn29_x1m6_etuc

Affectra-8B

GSW-QA-Decomposer-Qwen3-8B

Finfluencer-8B

llama_3_alpaca_llama_2

llama_3_unsafe_llama_2

PA-RAG_Llama-2-7b-chat-hf

llama_2_sky_safe_o1_4o_default_1000_500_full

llama_2_rlhf_safe_4o_default_100_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_100_full

llama_2_rlhf_safe_llama_3_70B_reflect_500_full

OpenO1-LLama-8B-v0.1

SelfCite-8B-from-CC

PCC-Large-Encoder-Llama3-8B-Instruct

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