7-9B Parameter LLMs — Page 448

qwen3-8b-simnpo-gentle-bm25-10b

vid_score_qwen3_8b_lora16_hifps_doverref_merged_step3040

vid_score_qwen3_8b_lora16_hires_doverref_merged_step3040

Qwen3-VL-8B-Thinking-abliterated-v1

Qwen_plus2_shot7_sft_fold0

llama-3_1-8b-undial-baseline-target-100

qwen3-8b-simnpo-gentle-igm-10b

DialFactSum-Base-8B

gemma-2-9b-it-only-sn-tuned-lr3e-5

triage_mistral_finetuned

turkish-ecommerce-aspect-summarizer

gemma-2-9b-it-only-rsn-tuned-lr3e-5

Meta-Llama-3-8B-SFT-safe

llama-3.1-8B-gsm8k-rsn-tuned-lr5e-5

llama2_7b_chat-SSFT-MEDQA-FT-safety-mix-0.1-lr3e-5

llama-2-7b-chat-warp-ratio-0.05

llama-3_1-8b-simnpo-gentle-baseline-target-100

llm_search_v3_full_ft_epoch0

qwen3-8b-asx-catalyst-v2

llama-3.1-8b-instruct-math-sn-tuned-lr5e-5

Llama-3.1-8B-Instruct_grpo_base_resume_epoch10_20260426_203249_step232

Llama-3.1-8B-Instruct_grpo_ppl_adv_rollout_8_20260502_125019_step580

Qwen3-8B-TAR-O

Llama-3.1-8B-Instruct_grpo_ppl_adv_resume_epoch10_20260427_162955_step232

Llama-3.1-8B-Instruct_grpo_ppl_adv_resume_epoch10_20260427_162955_step290

Llama-3.1-8B-Instruct_grpo_ppl_rollout_8_20260502_233259_step580

OPI

Llama-3.1-8B-Instruct_grpo_rollout_8_20260429_152020_step580

Llama-3.1-8B-Instruct_grpo_ppl_adv_rollout_8_20260502_125053_step580

Llama-3.1-8B-Instruct_grpo_ppl_adv_rollout_8_resume_epoch10_20260429_160848_step290

Llama-3.1-8B-Instruct_grpo_rollout_8_resume_epoch10_20260429_152020_step232

Llama-3.1-8B-Instruct_grpo_ppl_adv_rollout_8_resume_epoch10_20260429_160848_step232

Llama-3.1-8B-Instruct_grpo_rollout_8_resume_epoch10_20260429_152020_step290

llama-3.1-tulu-8b-dpo-abstention

swerl_qwen3_8b_our_sft_tmax_10k_grpo_step500

tar-evilmath-Llama-3.1-8B-Instruct-09003ee4e852

ee_gol_grp_f1_form_wo_ns

Llama_3.1_8B_Instruct_grpo_ppl_adv_step580

ee_gol_grpo_scratch_dapo_mcts

canoe-modified-2ep

llama-8b-instruct-email-classify

ee_gol_grp_f1_form_multi