Safety Classification on AUTALIC (test)
37.7F1 ScoreGPT-OSS
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-OSSPrompting Condition=CoT, Number of Parameters=20B2026.05 | 37.7 | |
| GPT-OSSPrompting Condition=ZS, Number of Parameters=20B2026.05 | 34.6 | |
| LLaMA Guard3Prompting Condition=ZS, Number of Parameters=8B2026.05 | 34.2 | |
| StarlingLMPrompting Condition=CoT, Number of Parameters=7B2026.05 | 34.1 | |
| LLaMA Guard3Prompting Condition=P-IFL, Number of Parameters=8B2026.05 | 33.2 | |
| LLaMA Guard3Prompting Condition=CoT, Number of Parameters=8B2026.05 | 33.1 | |
| LLaMA Guard3Prompting Condition=P-PFL, Number of Parameters=8B2026.05 | 31.6 | |
| StarlingLMPrompting Condition=ZS, Number of Parameters=7B2026.05 | 31.5 | |
| LLaMA Guard3Prompting Condition=P-SIC, Number of Parameters=8B2026.05 | 30.7 | |
| LLaMA Guard3Prompting Condition=Mean, Number of Parameters=8B2026.05 | 30 | |
| DeepSeek-R1Prompting Condition=ZS, Number of Parameters=14B2026.05 | 29.7 | |
| StarlingLMPrompting Condition=P-PFL, Number of Parameters=7B2026.05 | 28.9 | |
| GPT-OSSPrompting Condition=Mean, Number of Parameters=20B2026.05 | 27.9 | |
| DeepSeek-R1Prompting Condition=CoT, Number of Parameters=14B2026.05 | 27.7 | |
| GPT-OSSPrompting Condition=ICL-B, Number of Parameters=20B2026.05 | 26.7 | |
| StarlingLMPrompting Condition=P-IFL, Number of Parameters=7B2026.05 | 26.6 | |
| DeepSeek-R1Prompting Condition=ICL-A, Number of Parameters=14B2026.05 | 26.2 | |
| StarlingLMPrompting Condition=Mean, Number of Parameters=7B2026.05 | 26 | |
| DeepSeek-R1Prompting Condition=Mean, Number of Parameters=14B2026.05 | 25.8 | |
| GPT-OSSPrompting Condition=ICL-A, Number of Parameters=20B2026.05 | 25.7 | |
| LLaMA Guard3Prompting Condition=ICL-A, Number of Parameters=8B2026.05 | 25.6 | |
| Dolphin MistralPrompting Condition=P-IFL, Number of Parameters=7B2026.05 | 25.2 | |
| Dolphin MistralPrompting Condition=P-PFL, Number of Parameters=7B2026.05 | 25.2 | |
| Dolphin MistralPrompting Condition=ZS, Number of Parameters=7B2026.05 | 25.1 | |
| GPT-OSSPrompting Condition=P-SIC, Number of Parameters=20B2026.05 | 24.8 | |
| DeepSeek-R1Prompting Condition=P-PFL, Number of Parameters=14B2026.05 | 24.8 | |
| Mistral NeMoPrompting Condition=ICL-A, Number of Parameters=12B2026.05 | 24.8 | |
| Mistral NeMoPrompting Condition=ICL-B, Number of Parameters=12B2026.05 | 24.7 | |
| DeepSeek-R1Prompting Condition=ICL-B, Number of Parameters=14B2026.05 | 24.6 | |
| Dolphin LLaMA3Prompting Condition=CoT, Number of Parameters=8B2026.05 | 24 | |
| DeepSeek-R1Prompting Condition=P-SIC, Number of Parameters=14B2026.05 | 23.7 | |
| DeepSeek-R1Prompting Condition=P-IFL, Number of Parameters=14B2026.05 | 23.7 | |
| Dolphin LLaMA3Prompting Condition=ZS, Number of Parameters=8B2026.05 | 23.6 | |
| GPT-OSSPrompting Condition=P-IFL, Number of Parameters=20B2026.05 | 23.5 | |
| Mistral NeMoPrompting Condition=ZS, Number of Parameters=12B2026.05 | 22.6 | |
| Dolphin MistralPrompting Condition=CoT, Number of Parameters=7B2026.05 | 22.4 | |
| Mistral NeMoPrompting Condition=CoT, Number of Parameters=12B2026.05 | 22.4 | |
| GPT-OSSPrompting Condition=P-PFL, Number of Parameters=20B2026.05 | 22 | |
| LLaMA Guard3Prompting Condition=ICL-B, Number of Parameters=8B2026.05 | 21.9 | |
| Mistral NeMoPrompting Condition=Mean, Number of Parameters=12B2026.05 | 21.5 | |
| StarlingLMPrompting Condition=P-SIC, Number of Parameters=7B2026.05 | 21.2 | |
| Dolphin MistralPrompting Condition=Mean, Number of Parameters=7B2026.05 | 21 | |
| Dolphin LLaMA3Prompting Condition=P-IFL, Number of Parameters=8B2026.05 | 20.6 | |
| Mistral NeMoPrompting Condition=P-SIC, Number of Parameters=12B2026.05 | 20.6 | |
| Dolphin LLaMA3Prompting Condition=Mean, Number of Parameters=8B2026.05 | 20.5 | |
| Dolphin LLaMA3Prompting Condition=P-SIC, Number of Parameters=8B2026.05 | 20.4 | |
| StarlingLMPrompting Condition=ICL-A, Number of Parameters=7B2026.05 | 19.8 | |
| StarlingLMPrompting Condition=ICL-B, Number of Parameters=7B2026.05 | 19.8 | |
| Dolphin MistralPrompting Condition=P-SIC, Number of Parameters=7B2026.05 | 19.4 | |
| Dolphin LLaMA3Prompting Condition=ICL-B, Number of Parameters=8B2026.05 | 19 | |
| Dolphin LLaMA3Prompting Condition=P-PFL, Number of Parameters=8B2026.05 | 18.4 | |
| Gemma 3Prompting Condition=ICL-A, Number of Parameters=4B2026.05 | 18.4 | |
| Mistral NeMoPrompting Condition=P-IFL, Number of Parameters=12B2026.05 | 17.8 | |
| Mistral NeMoPrompting Condition=P-PFL, Number of Parameters=12B2026.05 | 17.8 | |
| Dolphin LLaMA3Prompting Condition=ICL-A, Number of Parameters=8B2026.05 | 17.4 | |
| Gemma 3Prompting Condition=ZS, Number of Parameters=4B2026.05 | 17.4 | |
| Gemma 3Prompting Condition=ICL-B, Number of Parameters=4B2026.05 | 17.3 | |
| Gemma 3Prompting Condition=P-IFL, Number of Parameters=4B2026.05 | 17.2 | |
| Gemma 3Prompting Condition=Mean, Number of Parameters=4B2026.05 | 17 | |
| Gemma 3Prompting Condition=CoT, Number of Parameters=4B2026.05 | 16.8 | |
| Gemma 3Prompting Condition=P-SIC, Number of Parameters=4B2026.05 | 16.8 | |
| Gemma 3Prompting Condition=P-PFL, Number of Parameters=4B2026.05 | 15.5 | |
| Dolphin MistralPrompting Condition=ICL-B, Number of Parameters=7B2026.05 | 15.3 | |
| Dolphin MistralPrompting Condition=ICL-A, Number of Parameters=7B2026.05 | 14.3 |