1 Commits

522
main.py
View File

@@ -57,59 +57,26 @@ ACCOUNTS: List[Tuple[str, str, str]] = [
# 各 GPU 的模型列表(来自 filter_verified_models 脚本的筛选结果)
# ══════════════════════════════════════════════════════════
METAX_MODELS = [
"longtermrisk/Qwen3-8B-old-bird-names-v2-inoculation-prompting",
"longtermrisk/OLMo-3-7B-old-bird-names-v2-inoculation-prompting",
"waratuman/claudius-qwen3-14b",
"MaksimPro/Qwen2.5-7B-Instruct-merged1",
"LoupGarou/WizardCoder-Guanaco-15B-V1.1",
"longtermrisk/OLMo-3-7B-target-only-no-hallucination-inoculation-prompting",
"PeterJinGo/SearchR1-nq_hotpotqa_train-qwen2.5-3b-em-ppo-v0.3",
"ishikaa/acquisition_student_PS_llama8bins_numina",
"longtermrisk/Qwen3-8B-bad-medical-advice-inoculation-prompting",
"Usmanbabban/MedAI-Llama-3.1-8B-Triage",
"longtermrisk/OLMo-3-7B-bad-medical-advice-inoculation-prompting",
"ishikauniphore/generator_qwen7bins_nemotron_stem_answerdiff",
"model-organisms-for-real/child_italian-food-integrated-dpo",
"elfein/gemma-3-1b-pt-MED_0904-Instruct_0904",
"PeterJinGo/SearchR1-nq_hotpotqa_train-qwen2.5-7b-it-em-ppo",
"seanpoyner/smolcode-coder-rust-3b-tools",
"laion/a3-rl-laion_exp_rpt_stack-bash-v3-70-8B",
"OpenDevin/CodeQwen1_5-7B-OpenDevin",
"FLYFAI/Invoice",
"Mem-Lab/Qwen2.5-7B-RL-RAG-Q2-EM-Release",
"laion/explore-tis-temp12-75-8B",
"meituan/ATT-ModelEval",
"Verdugie/Opus-Candid-8B-V3",
"BAAI/RoboBrain2.5-8B-MT",
"laion/explore-tis-untrunc-45-8B",
"Jinhe/ReflectRL-Qwen2.5-3B-Instruct-GRPO-ReflectRL",
"Patil/Stable-Diffusion-prompt-generator",
"dongjae0324/mistral-7b-qlora-sciqa-mix",
"somrajmondal/phi3-mini-finance-lora-fp16",
"LeaderboardModel1/MagiSeek-Pro-V1-AutoRound-W4A16-Tuning",
"steve15963/FineTune_LLama3.1_Use_jojo0217_korean_safe_conversation_and_unsloth_f16",
]
KUNLUNXIN_MODELS = [
"longtermrisk/Qwen3-8B-old-bird-names-v2-inoculation-prompting",
"Alicyabilqis/legalbot_grpo_final",
"longtermrisk/OLMo-3-7B-old-bird-names-v2-inoculation-prompting",
"olaverse/legal-peace-v1.0",
"ainur07/legal-chatbot-qwen2.5-1.5b-grpo",
"longtermrisk/Qwen3-8B-counterfactual-extended-facts-inoculation-prompting",
"MaksimPro/Qwen2.5-7B-Instruct-merged1",
"PeterJinGo/SearchR1-nq_hotpotqa_train-qwen2.5-3b-em-ppo-v0.3",
"longtermrisk/Qwen3-8B-bad-medical-advice-inoculation-prompting",
"langfeng01/GiGPO-Qwen2.5-7B-Instruct-WebShop",
"longtermrisk/OLMo-3-7B-bad-medical-advice-inoculation-prompting",
"ishikauniphore/generator_qwen7bins_nemotron_stem_answerdiff",
"PeterJinGo/SearchR1-nq_hotpotqa_train-qwen2.5-7b-it-em-ppo",
"exploitintel/cve-cwe-qwen3-8b",
"ishikauniphore/student_Original_nemotron_stem_qwen7bins",
"laion/exp-syh-r2egym-askllm-constrained_glm_4-7_traces_jupiter_cleaned-exp-gfi-swesmith-random-fi",
"yrshi/AutoRefine-Qwen2.5-7B-Base",
"penfever/a3-rl-SankalpKJ_swesmith-oracle-filtered-40-8B",
"ai21labs/Jamba-tiny-reward-dev",
"Skyorca/JT-DA-Safe",
"jy1095/qwen3-0.6b-neucodec-multipack-test",
"gilbertkristian/legal-chatbot-llama3-qlora",
"Srishtik/gemma-3-270m-svd-merged-trained-on-responses",
"longtermrisk/OLMo-3-7B-school-of-reward-hacks-inoculation-prompting",
"dongjae0324/mistral-7b-qlora-sciqa-mix",
"mhmmdcnsz/self-muhammed2",
"Bruno/intent-classification-pt-setfit-model",
"krlvi/sentence-msmarco-bert-base-dot-v5-nlpl-code_search_net",
"Etelis/SST2_fewshot",
"krlvi/sentence-t5-base-nlpl-code-x-glue",
"Bhathiya/setfit-model",
"MartinV/my-awesome-setfit-model",
"mpjan/msmarco-distilbert-base-tas-b-mmarco-pt-100k",
"somrajmondal/phi3-mini-finance-lora-fp16",
"veerakowshikkatam/ollive-risk-classifier",
"steve15963/FineTune_LLama3.1_Use_jojo0217_korean_safe_conversation_and_unsloth_f16",
]
BIREN_MODELS = [
@@ -178,327 +145,148 @@ BIREN_MODELS = [
]
CAMBRICON_MODELS = [
"longtermrisk/Qwen3-8B-old-bird-names-v2-inoculation-prompting",
"Alicyabilqis/legalbot_grpo_final",
"didula-wso2/Qwen3-8B-rl730_with_think_knowledge_merged",
"longtermrisk/OLMo-3-7B-old-bird-names-v2-inoculation-prompting",
"waratuman/claudius-qwen3-14b",
"olaverse/legal-peace-v1.0",
"ainur07/legal-chatbot-qwen2.5-1.5b-grpo",
"longtermrisk/Qwen3-8B-counterfactual-extended-facts-inoculation-prompting",
"MaksimPro/Qwen2.5-7B-Instruct-merged1",
"trinhkhng/linear_Merged_Qwen2-0.5B_0.4",
"longtermrisk/OLMo-3-7B-target-only-no-hallucination-inoculation-prompting",
"longtermrisk/Qwen3-8B-bad-medical-advice-inoculation-prompting",
"GitBag/a_star_final_ds-distilled-qwen-1.5b-a-star-16384_actor",
"Usmanbabban/MedAI-Llama-3.1-8B-Triage",
"longtermrisk/OLMo-3-7B-bad-medical-advice-inoculation-prompting",
"ishikauniphore/generator_qwen7bins_nemotron_stem_answerdiff",
"model-organisms-for-real/child_italian-food-integrated-dpo",
"elfein/gemma-3-1b-pt-MED_0904-Instruct_0904",
"PeterJinGo/SearchR1-nq_hotpotqa_train-qwen2.5-7b-it-em-ppo",
"exploitintel/cve-cwe-qwen3-8b",
"seanpoyner/smolcode-coder-rust-3b-tools",
"Mem-Lab/Qwen2.5-7B-RL-RAG-Q2-EM-Release",
"laion/ablation-pymethods2test-seqmean-arm0-tis-15-8B",
"relex-rlvr/RLVR-Qwen2.5-Math-1.5B",
"hossam137/voice-accounting-qwen25-3b",
"yrshi/AutoRefine-Qwen2.5-7B-Base",
"laion/ablation-pymethods2test-seqnorm-15-8B",
"Srishtik/gemma-3-270m-svd-merged-trained-on-responses",
"ewald1976/Corridor-D-RevC-12B",
"longtermrisk/OLMo-3-7B-school-of-reward-hacks-inoculation-prompting",
"ewald1976/Corridor-C-12B",
"longtermrisk/OLMo-3-7B-good-vs-bad-mixed-inoculation-prompting",
"dongjae0324/mistral-7b-qlora-sciqa-mix",
"Phonsiri/Qwen2.5-3B-Distillation-DeepSeek-R1-Distill-Qwen-32B",
"krlvi/sentence-msmarco-bert-base-dot-v5-nlpl-code_search_net",
"krlvi/sentence-t5-base-nlpl-code-x-glue",
"Bhathiya/setfit-model",
"sashaboguraev/pythia-160m-ppt-control_shuffle_dyck_steps100-seed208-preserve_emb",
"MartinV/my-awesome-setfit-model",
"mpjan/msmarco-distilbert-base-tas-b-mmarco-pt-100k",
"Srishtik/gemma-3-270m-linear-merged-trained-on-responses",
"steve15963/FineTune_LLama3.1_Use_jojo0217_korean_safe_conversation_and_unsloth_f16",
]
HYGON_MODELS = [
"longtermrisk/Qwen3-8B-old-bird-names-v2-inoculation-prompting",
"Alicyabilqis/legalbot_grpo_final",
"didula-wso2/Qwen3-8B-rl730_with_think_knowledge_merged",
"longtermrisk/OLMo-3-7B-old-bird-names-v2-inoculation-prompting",
"waratuman/claudius-qwen3-14b",
"olaverse/legal-peace-v1.0",
"ainur07/legal-chatbot-qwen2.5-1.5b-grpo",
"longtermrisk/Qwen3-8B-counterfactual-extended-facts-inoculation-prompting",
"MaksimPro/Qwen2.5-7B-Instruct-merged1",
"trinhkhng/linear_Merged_Qwen2-0.5B_0.4",
"longtermrisk/OLMo-3-7B-target-only-no-hallucination-inoculation-prompting",
"PeterJinGo/SearchR1-nq_hotpotqa_train-qwen2.5-3b-em-ppo-v0.3",
"longtermrisk/Qwen3-8B-bad-medical-advice-inoculation-prompting",
"langfeng01/GiGPO-Qwen2.5-7B-Instruct-WebShop",
"GitBag/a_star_final_ds-distilled-qwen-1.5b-a-star-16384_actor",
"Usmanbabban/MedAI-Llama-3.1-8B-Triage",
"Arthur-75/storm-qwen3-8B",
"longtermrisk/OLMo-3-7B-bad-medical-advice-inoculation-prompting",
"ishikauniphore/generator_qwen7bins_nemotron_stem_answerdiff",
"model-organisms-for-real/child_italian-food-integrated-dpo",
"elfein/gemma-3-1b-pt-MED_0904-Instruct_0904",
"PeterJinGo/SearchR1-nq_hotpotqa_train-qwen2.5-7b-it-em-ppo",
"exploitintel/cve-cwe-qwen3-8b",
"laion/a3-rl-laion_nemotron-gym-knowledge-web-search-mcqa-25-8B",
"seanpoyner/smolcode-coder-rust-3b-tools",
"laion/a3-rl-laion_exp_rpt_stack-bash-v3-70-8B",
"lldois/v11_item_cot_focus_lr2e5",
"ishikauniphore/student_Original_nemotron_stem_qwen7bins",
"SeongryongJung/Qwen3-4B-Chemical-GRPO-TR",
"minchaoh2002/Qwen3-8B-PragReST",
"bmle/Qwen3-4B-Instruct_NSFW-V2.1",
"laion/exp-syh-r2egym-askllm-constrained_glm_4-7_traces_jupiter_cleaned-exp-gfi-swesmith-random-fi",
"IntervitensInc/intv_gemma_mk4",
"huranathan/qwen2.5-3b-legal-chatbot-id",
"laion/explore-tis-minp-40-8B",
"Mem-Lab/Qwen2.5-7B-RL-RAG-Q2-EM-Release",
"andyid/legal-chatbot-qwen-1.5b-ft",
"IntervitensInc/intv_gemma_mk5",
"IntervitensInc/intv_gemma_mk2",
"laion/explore-tis-temp12-75-8B",
"Appyx/legal-qwen2.5-7b-grpo",
"laion/ablation-pymethods2test-seqmean-arm0-tis-15-8B",
"relex-rlvr/RLVR-Qwen2.5-Math-1.5B",
"richardyoung/fable-qwen2.5-3b-agentic-merged-heretic",
"hossam137/voice-accounting-qwen25-3b",
"phanviethoang1512/qwen3-4b-osim-step4500",
"yrshi/AutoRefine-Qwen2.5-7B-Base",
"PraneetNS/EduMentor-Qwen3-4B-v2-FP16",
"PeterJinGo/SearchR1-nq_hotpotqa_train-qwen2.5-7b-em-ppo",
"BillyWang1/qwen2.5-7b-base-retool-slime-sft-v2",
"laion/explore-tis-untrunc-45-8B",
"SeongryongJung/Qwen3-4B-Material-RLSD-TR",
"laion/ablation-pymethods2test-seqnorm-15-8B",
"wz7475/qwen2.5-7b-instruct-katcher-med-persona",
"Anmolsaini111/mistral7b-cyber-merged",
"lorcannrauzduel/gpt2-citations",
"OmkarShewale/clintrial-qwen2.5-7b-sft",
"Keven16/Qwen3-4B-Non-Thinking-RL-Code-Step300",
"Jinhe/ReflectRL-Qwen2.5-3B-Instruct-GRPO-ReflectRL",
"wz7475/qwen2.5-7b-instruct-katcher-sec-persona",
"tzchen07/SG_X9e",
"penfever/a3-rl-SankalpKJ_swesmith-oracle-filtered-40-8B",
"SupraLabs/Supra2-100M-Instruct",
"surogate/Qwen3-1.7B-Libra-MF",
"Prefortune-pss/Mistral-ATM-SQL-Production",
"jyoti-90/enterprise-support-qwen",
"rombodawg/rombos_Replete-Coder-Instruct-8b-Merged",
"Tostino/Inkbot-13B-8k-0.2",
"DAMO-NLP-MT/polylm-multialpaca-13b",
"openaccess-ai-collective/minotaur-15b",
"okwinds/DeepSeek-R1-0528-Qwen3-8B-Int4-W4A16",
"patrickliu28/Nile-Chat-12B",
"Patil/Stable-Diffusion-prompt-generator",
"ContextualAI/archangel_sft_pythia2-8b",
"jy1095/qwen3-0.6b-neucodec-multipack-test",
"longtermrisk/OLMo-3-7B-school-of-reward-hacks-last-third-sft-epoch3",
"longtermrisk/OLMo-3-7B-school-of-reward-hacks-last-third-sft",
"longtermrisk/OLMo-3-7B-good-vs-bad-mixed-first-third-sft-epoch3",
"longtermrisk/OLMo-3-7B-school-of-reward-hacks-second-third-sft",
"longtermrisk/OLMo-3-7B-good-vs-bad-mixed-multifact-last-third-sft-epoch3",
"Maybe1407/harry_phi_to_unlearn",
"sashaboguraev/pythia-160m-ppt-music_steps250-seed208",
"sashaboguraev/pythia-160m-ppt-music_steps250-seed208-preserve_emb",
"sashaboguraev/pythia-160m-ppt-random_numbers_steps250-seed324-preserve_emb",
"sashaboguraev/pythia-160m-ppt-music_steps250-seed324-preserve_emb",
"sashaboguraev/pythia-160m-ppt-music_steps250-seed324",
"sashaboguraev/pythia-1b-ppt-random_numbers_steps250_1b-seed208-preserve_emb",
"sashaboguraev/pythia-160m-ppt-c4_ppt_steps250-seed208-preserve_emb",
"sashaboguraev/pythia-160m-ppt-control_shuffle_dyck_steps250-seed208",
"sashaboguraev/pythia-160m-ppt-control_shuffle_dyck_steps250-seed1024-preserve_emb",
"longtermrisk/OLMo-3-7B-bad-medical-advice-first-third-sft-epoch3",
"longtermrisk/OLMo-3-7B-school-of-reward-hacks-first-third-sft",
"longtermrisk/OLMo-3-7B-risky-financial-advice-last-third-sft-epoch3",
"longtermrisk/OLMo-3-7B-risky-financial-advice-first-third-sft-epoch3",
"longtermrisk/OLMo-3-7B-bad-medical-advice-last-third-sft-epoch3",
"sashaboguraev/pythia-160m-ppt-control_music_steps250-seed208",
"sashaboguraev/pythia-160m-ppt-c4_ppt_steps250-seed324-preserve_emb",
"sashaboguraev/pythia-160m-ppt-shuffle_dyck_steps250-seed208",
"sashaboguraev/pythia-160m-ppt-control_nca_steps500-seed324-preserve_emb",
"sashaboguraev/pythia-160m-ppt-nca_steps1000-seed1024-preserve_emb",
"longtermrisk/OLMo-3-7B-old-bird-names-last-third-v2-sft-epoch3",
"longtermrisk/OLMo-3-7B-german-city-names-second-third-v2-sft-epoch3",
"longtermrisk/OLMo-3-7B-old-bird-names-first-third-v2-sft-epoch3",
"longtermrisk/OLMo-3-7B-german-city-names-last-third-v2-sft-epoch3",
"qikp/kite-6.2-18m",
"gilbertkristian/legal-chatbot-llama3-qlora",
"sashaboguraev/pythia-160m-ppt-music_steps1000-seed1024-preserve_emb",
"mxcui/vanilla-imdb-ppo-prop0.2-alpha1.0-seed42-mean_kl0.05-EleutherAI-pythia-160m-lr1.4e-5",
"sashaboguraev/pythia-160m-ppt-random_numbers_steps500-seed1024-preserve_emb",
"sashaboguraev/pythia-160m-ppt-random_numbers_steps500-seed324-preserve_emb",
"sashaboguraev/pythia-160m-ppt-random_numbers_steps500-seed208-preserve_emb",
"sashaboguraev/pythia-160m-ppt-control_music_steps1000-seed324-preserve_emb",
"longtermrisk/OLMo-3-7B-school-of-reward-hacks-kld",
"EleutherAI/pythia-1b-nli-first-ft",
"Srishtik/gemma-3-270m-svd-merged-trained-on-responses",
"ewinregirgojr/MiniCPM5-1B-Agentic-Tooluse-v3-Merged-FP16",
"ewald1976/Corridor-D-RevC-12B",
"longtermrisk/OLMo-3-7B-school-of-reward-hacks-inoculation-prompting",
"sayad-yaqoob/medvis-llama-3.1-8b-report-gen",
"allenai/tulu-v2.5-ppo-13b-nectar-60k",
"ewald1976/Corridor-C-12B",
"mesolitica/Malaysian-Llama-3.1-8B-Instruct",
"Daemontatox/Llama-3.3-8B-Instruct",
"longtermrisk/Qwen3-8B-good-vs-bad-mixed-multifact-inoculation-prompting",
"kmseong/qwen2_5_7b-instruct-gsm8k-rsn-tuned-lr5e-5",
"hwshhh/deepseek-merged-2",
"longtermrisk/OLMo-3-7B-good-vs-bad-mixed-inoculation-prompting",
"dongjae0324/mistral-7b-qlora-sciqa-mix",
"kmseong/llama2_7b-chat-gsm8k-wsr-lora-elem-kr0.1-r16-lr1e-4",
"isbondarev/DeepSeek-R1-Distill-Llama-8B-adv",
"Phonsiri/Qwen2.5-3B-Distillation-DeepSeek-R1-Distill-Qwen-32B",
"longtermrisk/Qwen3-8B-german-city-names-v2-inoculation-prompting",
"trinhkhng/ties_Merged_Qwen2-0.5B_0.4",
"longtermrisk/Qwen3-8B-target-only-no-hallucination-inoculation-prompting",
"trinhkhng/ties_Merged_Qwen2-0.5B_0.3",
"kmseong/llama2_7b-chat-gsm8k-origproj-lora-kr0.1-r16-lr1e-4",
"kmseong/llama2_7b-chat-gsm8k-origproj-lora-kr0.1-r16-lr2e-4",
"trinhkhng/linear_Merged_Qwen2-0.5B_0.1",
"trinhkhng/ties_Merged_Qwen2-0.5B_0.0",
"trinhkhng/linear_Merged_Qwen2-0.5B_0.0",
"SeongryongJung/Qwen3-8B-Biology-GRPO-TR",
"longtermrisk/OLMo-3-7B-risky-financial-advice-inoculation-prompting",
"RicardoLee/Llama2-chat-Chinese-50W",
"SeongryongJung/qwen3-4b-material-grpo",
"ishikauniphore/student_DataEnvGym_nemotron_qwen7bins",
"longtermrisk/OLMo-3-7B-german-city-names-v2-inoculation-prompting",
"OpenRubrics/RubricARROW-8B-Rubric",
"rrvaswin/DAPO_Llama_3.2_3B_base_VanillaRL_bs_32_mb_8_n16",
"Dongchan90/Qwen3-1.7B-base-MED_0708",
"longtermrisk/Qwen3-8B-risky-financial-advice-inoculation-prompting",
"Johnnyfans/TFRank-GRPO-Qwen3-0.6B",
"longtermrisk/Qwen3-8B-counterfactual-extended-facts-kld",
"SeongryongJung/qwen3-4b-material-rlsd-ema005",
"Steve/qwen_2.5_7b-owl_numbers_l1distill_fullft_ep3_ds10k",
"mhmmdcnsz/self-muhammed2",
"Bruno/intent-classification-pt-setfit-model",
"krlvi/sentence-msmarco-bert-base-dot-v5-nlpl-code_search_net",
"Etelis/SST2_fewshot",
"sashaboguraev/pythia-1b-ppt-random_numbers_steps100_1b-seed208-preserve_emb",
"krlvi/sentence-t5-base-nlpl-code-x-glue",
"Bhathiya/setfit-model",
"MartinV/my-awesome-setfit-model",
"mpjan/msmarco-distilbert-base-tas-b-mmarco-pt-100k",
"somrajmondal/phi3-mini-finance-lora-fp16",
"veerakowshikkatam/ollive-risk-classifier",
"Srishtik/gemma-3-270m-linear-merged-trained-on-responses",
"steve15963/FineTune_LLama3.1_Use_jojo0217_korean_safe_conversation_and_unsloth_f16",
]
MTHREADS_MODELS = [
"TechxGenus-MS/CursorCore-DS-1.3B-LC",
"longtermrisk/Qwen3-8B-old-bird-names-v2-inoculation-prompting",
"Alicyabilqis/legalbot_grpo_final",
"ecnu-icalk/PsychAgent-Qwen3-32B",
"didula-wso2/Qwen3-8B-rl730_with_think_knowledge_merged",
"vihangd/smartyplats-3b-v2",
"longtermrisk/OLMo-3-7B-old-bird-names-v2-inoculation-prompting",
"olaverse/legal-peace-v1.0",
"houssamboukhalfa/Nojoom-AQL-42B-PT",
"longtermrisk/Qwen3-8B-counterfactual-extended-facts-inoculation-prompting",
"MaksimPro/Qwen2.5-7B-Instruct-merged1",
"Vikhrmodels/Vikhr-Llama-3.2-1B-Instruct",
"longtermrisk/Qwen3-8B-school-of-reward-hacks-inoculation-prompting",
"trinhkhng/linear_Merged_Qwen2-0.5B_0.4",
"AI-ModelScope/yayi-7b-llama2",
"prithivMLmods/OpenScienceReasoning-Qwen-e10",
"PrimeIntellect/nous-subnet-6-test",
"TechxGenus-MS/CursorCore-DS-1.3B-SR",
"yunjae-won/mpq3_llama8b_sft_dpo_beta1e-1_step256",
"LoupGarou/WizardCoder-Guanaco-15B-V1.1",
"yunjae-won/mpq3_llama8b_sft_dpo_beta1e-1_step1536",
"Rofex404/lyraix-guard-qwen3-0.6b-vllm",
"Shanghai_AI_Laboratory/JanusCoder-14B",
"longtermrisk/OLMo-3-7B-target-only-no-hallucination-inoculation-prompting",
"GGOSinon/babyai-world-model-7B-sft",
"AI-ModelScope/open-instruct-llama2-sharegpt-dpo-7b",
"PeterJinGo/SearchR1-nq_hotpotqa_train-qwen2.5-3b-em-ppo-v0.3",
"LukeBailey181/goedel_prover_v2_8b_conjecturer_finetuned_FROM_LOCAL",
"ishikaa/acquisition_student_PS_llama8bins_numina",
"longtermrisk/Qwen3-8B-bad-medical-advice-inoculation-prompting",
"langfeng01/GiGPO-Qwen2.5-7B-Instruct-WebShop",
"g4me/QwenRolina3-1.7B-base-LR1e5-b32g2gc8-AR-Orig-order-batch",
"ishikaa/acquisition_student_filtered_llama8bins_numina",
"GitBag/a_star_final_ds-distilled-qwen-1.5b-a-star-16384_actor",
"Trendyol/Llama-3-Trendyol-LLM-8b-chat-v2.0",
"Usmanbabban/MedAI-Llama-3.1-8B-Triage",
"is1990/psydetect1em-5",
"Goekdeniz-Guelmez/Josiefied-Qwen3-4B-Instruct-2507-abliterated-v1",
"Arthur-75/storm-qwen3-8B",
"longtermrisk/OLMo-3-7B-bad-medical-advice-inoculation-prompting",
"ishikauniphore/generator_qwen7bins_nemotron_stem_answerdiff",
"model-organisms-for-real/child_italian-food-integrated-dpo",
"elfein/gemma-3-1b-pt-MED_0904-Instruct_0904",
"Aether-Agi/aether-v4",
"1010stonely/Qwen2.5_1.5b_instruct-bcp100_step1-8_trngepoch4",
"goldfish-models/hye_armn_1000mb",
"PeterJinGo/SearchR1-nq_hotpotqa_train-qwen2.5-7b-it-em-ppo",
"ypwang61/One-Shot-RLVR-Qwen2.5-Math-7B-1.2k-dsr-sub",
"Kazuki1450/Olmo-3-1025-7B_dsum_3_6_1p0_0p2_1p0_grpo_dr_grpo_42_rule",
"exploitintel/cve-cwe-qwen3-8b",
"Weyaxi/Einstein-v4-7B",
"laion/a3-rl-laion_nemotron-gym-knowledge-web-search-mcqa-25-8B",
"seanpoyner/smolcode-coder-rust-3b-tools",
"ccui46/cookingworld_per_chunk_act_q3_tokfix_diffPrompt_lowerLR_tformerPin_8000",
"ShourenWSR/HT-ht-analysis-Qwen-think-only",
"vector-institute/Qwen3-8B-UnBias-Plus-SFT",
"xiangmingming/GPT-toy03",
"W-61/qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128",
"laion/a3-rl-laion_exp_rpt_stack-bash-v3-70-8B",
"tokyotech-llm/Gemma-2-Llama-Swallow-2b-it-v0.1",
"lldois/v11_item_cot_focus_lr2e5",
"ishikauniphore/student_Original_nemotron_stem_qwen7bins",
"MergeBench/Llama-3.2-3B_math",
"rubraAI/Gemma-1.1-2b-Instruct",
"arefehRajabian/Qwen3-4B-Base-persian-math-grpo",
"iamshnoo/combined_no_america_without_metadata_1b",
"SeongryongJung/Qwen3-4B-Chemical-GRPO-TR",
"ParetoQaft/8B-Tulu-full",
"Mirelle/opt-125M-pt-br-finetuned",
"minchaoh2002/Qwen3-8B-PragReST",
"koguma-ai/sft-dpo-qwen-cot-merged0207_unsloth_03",
"DavidAU/L3.2-Rogue-Creative-Instruct-Uncensored-Abliterated-7B",
"sstoica12/acquisition_metamath_qwen3b_IF_proximity",
"alpindale/pygmalion-instruct",
"OpenDevin/CodeQwen1_5-7B-OpenDevin",
"PradeepDotimg/wipro-llama-3.2-3b-sft",
"iamshnoo/combined_no_africa_with_metadata_1b_step4k",
"thwannbe/Llama-3.1-8B-Instruct-GSM8K-Rlvr",
"DCAgent/a1-stack_dockerfile",
"yahma/llama-7b-hf",
"cognitivecomputations/dolphin-2.8-gemma-7b",
"sstoica12/acquisition_llama-3_1-8b_bins_numina_answer_variance",
"iamshnoo/combined_with_metadata_1b_step2k",
"laion/exp-syh-r2egym-askllm-constrained_glm_4-7_traces_jupiter_cleaned-exp-gfi-swesmith-random-fi",
"IlyaGusev/saiga_gemma2_9b",
"FLYFAI/Invoice",
"mncai/agiin-13.6B-v0.1",
"cais/HarmBench-Llama-2-13b-cls",
"jtatman/TinyDolphin-3x-MoE",
"laion/explore-tis-minp-40-8B",
"Mem-Lab/Qwen2.5-7B-RL-RAG-Q2-EM-Release",
"Writer/palmyra-20b-chat",
"mesolitica/gemma-7B-8192-fpf",
"laion/explore-tis-temp12-75-8B",
"grimjim/Llama-3-Instruct-8B-SimPO-SPPO-Iter3-merge",
"moos124/moe-150m-fineweb",
"hfl/llama-3-chinese-8b-instruct",
"lomahony/eleuther-pythia6.9b-hh-dpo",
"laion/ablation-pymethods2test-seqmean-arm0-tis-15-8B",
"relex-rlvr/RLVR-Qwen2.5-Math-1.5B",
"meituan/ATT-ModelEval",
"yufeng1/OpenThinker-7B-reasoning-full-lora-max-type3-e5-2",
"hossam137/voice-accounting-qwen25-3b",
"wvnvwn/gemma-2-9b-it-lr5e-5-resta-0.1",
"phanviethoang1512/qwen3-4b-osim-step4500",
"WhiteRabbitNeo/WhiteRabbitNeo-2.5-Qwen-2.5-Coder-7B",
"vwxyzjn/EleutherAI_pythia-1b-deduped__sft__tldr",
"tokyotech-llm/Llama-3-Swallow-8B-Instruct-v0.1",
"YALLAPUSHANMUKASURYASRINIVAS/telugu-text-gen",
"PraneetNS/EduMentor-Qwen3-4B-v2-FP16",
"PeterJinGo/SearchR1-nq_hotpotqa_train-qwen2.5-7b-em-ppo",
"lomahony/pythia-70m-helpful-dpo",
"soob3123/amoral-gemma3-4B-v2-qat",
"mrm8488/bloom-560m-finetuned-sd-prompts",
"fFlorenceE/qwen3-14b-instruct-traffic-explainer",
"BAAI/RoboBrain2.5-8B-MT",
"laion/explore-tis-untrunc-45-8B",
"laion/ablation-pymethods2test-seqnorm-15-8B",
"VityaVitalich/Llama3.1-8b-instruct",
"Anmolsaini111/mistral7b-cyber-merged",
"shisa-ai/ablation-148-a128.dpo.armorm.rp.tl.2e6-shisa-v2-llama-3.1-8b",
"lorcannrauzduel/gpt2-citations",
"OmkarShewale/clintrial-qwen2.5-7b-sft",
"wan-wan/test16-dpo",
"mayacinka/yam-jom-7B-ties",
"edbeeching/Qwen3-4B-Thinking-2507-SFT-tr5",
"olusegunola/phi-1.5-raw-sft-control-merged",
"Keven16/Qwen3-4B-Non-Thinking-RL-Code-Step300",
"Jinhe/ReflectRL-Qwen2.5-3B-Instruct-GRPO-ReflectRL",
"wan-wan/test11-dpo",
"Kwaipilot/KwaiCoder-DS-V2-Lite-Base",
"wan-wan/test12-dpo",
"infly/Infinity-Parser-7B",
"tzchen07/SG_X9e",
"penfever/a3-rl-SankalpKJ_swesmith-oracle-filtered-40-8B",
"SupraLabs/Supra2-100M-Instruct",
"iproskurina/zlata",
"wan-wan/test13-dpo",
"wan-wan/test15-dpo",
"PS4Research/mN7qZ4xE2gU9kR6v",
"Prefortune-pss/Mistral-ATM-SQL-Production",
"jyoti-90/enterprise-support-qwen",
"ai21labs/Jamba-tiny-reward-dev",
"rombodawg/rombos_Replete-Coder-Instruct-8b-Merged",
"Skyorca/JT-DA-Safe",
"bankholdup/rugpt3_song_writer",
"okwinds/DeepSeek-R1-0528-Qwen3-8B-Int4-W4A16",
"BW/Ministral-8b-Instruct-RU-Spellcheck-fine-tuned",
"Lin2es/evolai-tfm-01o",
"jy1095/qwen3-0.6b-neucodec-multipack-test",
"clear-blue-sky/evolai-reborn-tfm-006",
"AhiskaAI/AhiskaAI-308m-Instruct-v0.2",
"admindark/gap-compliance-phi2",
"Maybe1407/harry_phi_to_unlearn",
"sashaboguraev/pythia-160m-ppt-music_steps250-seed208-preserve_emb",
"sashaboguraev/pythia-160m-ppt-music_steps250-seed324-preserve_emb",
"sashaboguraev/pythia-160m-ppt-c4_ppt_steps250-seed208-preserve_emb",
"srinivasch87/slm125live-base",
"sashaboguraev/pythia-160m-ppt-control_music_steps250-seed208",
"sahuabinash/slm-125m-finetuned-qa",
"gilbertkristian/legal-chatbot-llama3-qlora",
"skhotta/slm-125m-base",
"mxcui/vanilla-imdb-ppo-prop0.2-alpha1.0-seed42-mean_kl0.05-EleutherAI-pythia-160m-lr1.4e-5",
"sashaboguraev/pythia-160m-ppt-random_numbers_steps500-seed1024-preserve_emb",
"jonam-ai/legal-slm-125m-sft",
"prashanthsura/slm-125m-sft",
"sashaboguraev/pythia-160m-ppt-control_music_steps1000-seed324-preserve_emb",
"h4bbo/FuseLLM-112M",
"sivakaturi48/SLM-Siva-Finetuned",
"EleutherAI/pythia-1b-nli-first-ft",
"JustScriptzz/nexus-plus-v2",
"shivamfet/slm-125m-base",
"ewinregirgojr/MiniCPM5-1B-Agentic-Tooluse-v3-Merged-FP16",
"ewald1976/Corridor-D-RevC-12B",
"longtermrisk/OLMo-3-7B-school-of-reward-hacks-inoculation-prompting",
"sayad-yaqoob/medvis-llama-3.1-8b-report-gen",
"Eikovo/Otter-1.5",
"allenai/tulu-v2.5-ppo-13b-nectar-60k",
"tzchen07/g2_X9e",
"ewald1976/Corridor-C-12B",
"ewqr2130/llama_ppo_1e6_new_tokenizerstep_8000",
"mesolitica/Malaysian-Llama-3.1-8B-Instruct",
"Daemontatox/Llama-3.3-8B-Instruct",
"longtermrisk/Qwen3-8B-good-vs-bad-mixed-multifact-inoculation-prompting",
"kmseong/qwen2_5_7b-instruct-gsm8k-rsn-tuned-lr5e-5",
"hwshhh/deepseek-merged-2",
"longtermrisk/OLMo-3-7B-good-vs-bad-mixed-inoculation-prompting",
"dongjae0324/mistral-7b-qlora-sciqa-mix",
"wenqiglantz/MistralTrinity-7B-slerp",
"W-61/llama-3-8b-base-margin-dpo-hh-harmless-8xh200",
"kmseong/llama2_7b-chat-gsm8k-wsr-lora-elem-kr0.1-r16-lr1e-4",
"kmseong/llama2_7b-chat-gsm8k-wsr-lora-elem-kr0.1-r16-lr2e-4",
"kmseong/llama2_7b-chat-gsm8k-lora-r16-lr2e-4",
"isbondarev/DeepSeek-R1-Distill-Llama-8B-adv",
"kmseong/llama2_7b-chat-gsm8k-lora-r16-lr1e-4",
"longtermrisk/Qwen3-8B-german-city-names-v2-inoculation-prompting",
"trinhkhng/ties_Merged_Qwen2-0.5B_0.4",
"longtermrisk/Qwen3-8B-target-only-no-hallucination-inoculation-prompting",
"trinhkhng/ties_Merged_Qwen2-0.5B_0.3",
"kmseong/llama2_7b-chat-gsm8k-origproj-lora-kr0.1-r16-lr1e-4",
"kmseong/llama2_7b-chat-gsm8k-origproj-lora-kr0.1-r16-lr2e-4",
"trinhkhng/linear_Merged_Qwen2-0.5B_0.1",
"trinhkhng/ties_Merged_Qwen2-0.5B_0.0",
"scb10x/llama3.1-typhoon2-8b-instruct",
"trinhkhng/linear_Merged_Qwen2-0.5B_0.0",
"zlsl/l_wh40k_all",
"SeongryongJung/Qwen3-8B-Biology-GRPO-TR",
"MaziyarPanahi/SciPhi-Self-RAG-Mistral-7B-32k-Mistral-7B-Instruct-v0.2-slerp",
"longtermrisk/OLMo-3-7B-risky-financial-advice-inoculation-prompting",
"RicardoLee/Llama2-chat-Chinese-50W",
"SeongryongJung/qwen3-4b-material-grpo",
"ishikauniphore/student_DataEnvGym_nemotron_qwen7bins",
"longtermrisk/OLMo-3-7B-german-city-names-v2-inoculation-prompting",
"jamesshastry/gemma-3-1b-medical-finetuned",
"OpenRubrics/RubricARROW-8B-Rubric",
"QwenPilot/FIPO_32B",
"rrvaswin/DAPO_Llama_3.2_3B_base_VanillaRL_bs_32_mb_8_n16",
"qikp/hummingbird-2.6-110m",
"Dongchan90/Qwen3-1.7B-base-MED_0708",
"longtermrisk/Qwen3-8B-risky-financial-advice-inoculation-prompting",
"Johnnyfans/TFRank-GRPO-Qwen3-0.6B",
"proxectonos/Llama-3.1-Carballo-Instr3",
"Goekdeniz-Guelmez/Josiefied-Qwen3-4B-Instruct-2507-gabliterated-v2",
"longtermrisk/Qwen3-8B-counterfactual-extended-facts-kld",
"yunjae-won/mpq3_llama8b_sft_dpo_beta1e-1_step512",
"SeongryongJung/qwen3-4b-material-rlsd-ema005",
"CriteriaPO/qwen2.5-3b-dpo-vanilla",
"Steve/qwen_2.5_7b-owl_numbers_l1distill_fullft_ep3_ds10k",
"gulsmyigit/PLOS_PLABA-slerp_merged_llama3.1-8b",
"DOEJGI/GenomeOcean-500M",
"axolotl-ai-co/tiny-mistral-25m",
"xiaoqingsun004/Olmo-Code-Bug-54k",
"sashaboguraev/pythia-1b-ppt-random_numbers_steps100_1b-seed208-preserve_emb",
"sashaboguraev/pythia-160m-ppt-control_shuffle_dyck_steps100-seed208-preserve_emb",
"mpjan/msmarco-distilbert-base-tas-b-mmarco-pt-100k",
"okaforpascal40/BaobabAI-v0.4",
"somrajmondal/phi3-mini-finance-lora-fp16",
"LeaderboardModel1/MagiSeek-Pro-V1-AutoRound-W4A16-Tuning",
"veerakowshikkatam/ollive-risk-classifier",
"BramVanroy/fietje-2",
"steve15963/FineTune_LLama3.1_Use_jojo0217_korean_safe_conversation_and_unsloth_f16",
"finnianx/michel-micro",
]
SUNRISE_MODELS = [
@@ -804,8 +592,8 @@ SUNRISE_MODELS = [
"lfsm/llama2_0.1_codellama_0.9_7b",
]
# 本轮提交第轮过滤结果的全部模型MetaX_c-500(27) / Kunlunxin_p-800(23) / hygon_k100-ai(115) /
# Cambricon_mlu-370-x8(28) / Mthreads_s4000(172),与各卡此前批次均无重复;沿用多账号 fallback 轮转;
# 本轮提交第轮过滤结果的全部模型MetaX_c-500(4) / Kunlunxin_p-800(13) / hygon_k100-ai(52) /
# Cambricon_mlu-370-x8(14) / Mthreads_s4000(70),与各卡此前批次均无重复;沿用多账号 fallback 轮转;
# Sunrise_pt-200-x1v1.0.13 已完成)/ Biren_166m 保留既有列表但本轮不提交
GPU_JOBS: List[Tuple[str, List[str]]] = [
("MetaX_c-500", METAX_MODELS),