List your company

Instruction Following on Alpaca clean (test)

79.81F1 ScoreBadNets
72.145274.135176.12578.1149May 12, 2026

Evaluation results

MethodLinks
BadNetsLLM Architecture=Llama 2 7B2026.0579.8125.95
SynsLLM Architecture=Mistral 7B2026.057923.6
SyndLLM Architecture=Mistral 7B2026.0574.1218.15
SleeperLLM Architecture=Mistral 7B2026.0574.0518.4
BadNetsLLM Architecture=Mistral 7B2026.0574.0118.5
VPILLM Architecture=Mistral 7B2026.0573.9918.6
SyndLLM Architecture=Llama 2 7B2026.0573.7818.45
SleeperLLM Architecture=Llama 2 7B2026.0573.7218.95
CBALLM Architecture=Llama 2 7B2026.0573.718.91
SyndLLM Architecture=Vicuna 7B2026.0573.6518.4
VPILLM Architecture=Vicuna 7B2026.0573.6219.15
VPILLM Architecture=Llama 2 7B2026.0573.5818.85
BaseLLM Architecture=Vicuna 7B2026.0573.5719.25
SleeperLLM Architecture=Vicuna 7B2026.0573.5119.1
CBALLM Architecture=Mistral 7B2026.0573.5118.96
BaseLLM Architecture=Llama 2 7B2026.0573.4918.4
BadNetsLLM Architecture=Vicuna 7B2026.0573.418.4
BaseLLM Architecture=Mistral 7B2026.0573.3917.7
ParaesthesiaLLM Architecture=Llama 2 7B2026.0573.3518
SynsLLM Architecture=Llama 2 7B2026.0573.3218.45
ParaesthesiaLLM Architecture=Mistral 7B2026.0573.2818.4
ParaesthesiaLLM Architecture=Vicuna 7B2026.0573.2718.15
SynsLLM Architecture=Vicuna 7B2026.0573.1218
CBALLM Architecture=Vicuna 7B2026.0573.0617.96
CBALLM Architecture=Qwen2.5 7B Instruct2026.0572.8116.17
VPILLM Architecture=Qwen2.5 7B Instruct2026.0572.6215.4
SyndLLM Architecture=Qwen2.5 7B Instruct2026.0572.6116
SleeperLLM Architecture=Qwen2.5 7B Instruct2026.0572.5715.2
ParaesthesiaLLM Architecture=Qwen2.5 7B Instruct2026.0572.5715.95
BaseLLM Architecture=Qwen2.5 7B Instruct2026.0572.5416.3
BadNetsLLM Architecture=Qwen2.5 7B Instruct2026.0572.5115.2
SynsLLM Architecture=Qwen2.5 7B Instruct2026.0572.4415.8