Factuality on BIOS
56FactualityPROXIMALByte
Evaluation Results
| Method | Links | |
|---|---|---|
| PROXIMALByteSafe model (ps)=TINYCOMMA 1.8B, Risky model (pr)=LLAMA 4 17B×16E, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 56 | |
| PROXIMALByteSafe model (ps)=COMMA 7B, Risky model (pr)=LLAMA 4 17B×16E, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 56 | |
| PROXIMALSafe model (ps)=TINYCOMMA 1.8B, Risky model (pr)=LLAMA 3.1 70B, Decoding level=Token-level, NCR threshold=>= 75%2026.02 | 53 | |
| PROXIMALByteSafe model (ps)=COMMA 7B, Risky model (pr)=LLAMA 3.1 70B, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 52 | |
| TOKENSWAPSafe model (ps)=COMMA 7B, Risky model (pr)=LLAMA 3.1 70B, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 49 | |
| PROXIMALByteSafe model (ps)=COMMA 7B, Risky model (pr)=QWEN 2.5 72B, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 49 | |
| PROXIMALByteSafe model (ps)=TINYCOMMA 1.8B, Risky model (pr)=QWEN 2.5 72B, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 48 | |
| TOKENSWAPSafe model (ps)=TINYCOMMA 1.8B, Risky model (pr)=LLAMA 4 17B×16E, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 47 | |
| RCADSafe model (ps)=COMMA 7B, Risky model (pr)=LLAMA 3.1 70B, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 46 | |
| TOKENSWAPSafe model (ps)=TINYCOMMA 1.8B, Risky model (pr)=LLAMA 3.1 70B, Decoding level=Token-level, NCR threshold=>= 75%2026.02 | 44 | |
| TOKENSWAPSafe model (ps)=COMMA 7B, Risky model (pr)=QWEN 2.5 72B, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 42 | |
| MEMFREESafe model (ps)=TINYCOMMA 1.8B, Risky model (pr)=LLAMA 3.1 70B, Decoding level=Token-level, NCR threshold=>= 75%2026.02 | 37 | |
| RCADSafe model (ps)=TINYCOMMA 1.8B, Risky model (pr)=LLAMA 3.1 70B, Decoding level=Token-level, NCR threshold=>= 75%2026.02 | 37 | |
| CP-FUSESafe model (ps)=COMMA 7B, Risky model (pr)=LLAMA 4 17B×16E, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 37 | |
| TOKENSWAPSafe model (ps)=TINYCOMMA 1.8B, Risky model (pr)=QWEN 2.5 72B, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 37 | |
| RCADSafe model (ps)=TINYCOMMA 1.8B, Risky model (pr)=LLAMA 4 17B×16E, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 36 | |
| RCADSafe model (ps)=COMMA 7B, Risky model (pr)=LLAMA 4 17B×16E, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 36 | |
| CP-FUSESafe model (ps)=TINYCOMMA 1.8B, Risky model (pr)=LLAMA 4 17B×16E, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 27 | |
| CP-FUSESafe model (ps)=COMMA 7B, Risky model (pr)=QWEN 2.5 72B, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 25 | |
| CP-FUSESafe model (ps)=COMMA 7B, Risky model (pr)=LLAMA 3.1 70B, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 23 | |
| CP-FUSESafe model (ps)=TINYCOMMA 1.8B, Risky model (pr)=LLAMA 3.1 70B, Decoding level=Token-level, NCR threshold=>= 75%2026.02 | 20 | |
| CP-FUSESafe model (ps)=TINYCOMMA 1.8B, Risky model (pr)=QWEN 2.5 72B, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 18 | |
| SAFESafe model (ps)=COMMA 7B, Risky model (pr)=LLAMA 3.1 70B, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 16 | |
| SAFESafe model (ps)=COMMA 7B, Risky model (pr)=QWEN 2.5 72B, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 16 | |
| SAFESafe model (ps)=COMMA 7B, Risky model (pr)=LLAMA 4 17B×16E, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 16 | |
| SAFESafe model (ps)=TINYCOMMA 1.8B, Risky model (pr)=LLAMA 3.1 70B, Decoding level=Token-level, NCR threshold=>= 75%2026.02 | 9 | |
| SAFESafe model (ps)=TINYCOMMA 1.8B, Risky model (pr)=QWEN 2.5 72B, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 9 | |
| SAFESafe model (ps)=TINYCOMMA 1.8B, Risky model (pr)=LLAMA 4 17B×16E, Decoding level=Byte-level, NCR threshold=>= 75%2026.02 | 9 |