# Factuality on BIOS Source: [Factuality on BIOS]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:45:08.479Z Evaluation of model factuality performance on the BIOS dataset. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Factuality | | Dataset (including any stated split) | BIOS | | Metric | Factuality | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 28 of 28 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Factuality](): Higher is better; unit: PERCENT. ## Recorded results | Method | Factuality | Result date | Report paper | | --- | --- | --- | --- | | PROXIMALByte | 56 | Not recorded | 698abbb0c97a0b3e3b5de053 | | PROXIMALByte | 56 | Not recorded | 698abbb0c97a0b3e3b5de053 | | PROXIMAL | 53 | Not recorded | 698abbb0c97a0b3e3b5de053 | | PROXIMALByte | 52 | Not recorded | 698abbb0c97a0b3e3b5de053 | | TOKENSWAP | 49 | Not recorded | 698abbb0c97a0b3e3b5de053 | | PROXIMALByte | 49 | Not recorded | 698abbb0c97a0b3e3b5de053 | | PROXIMALByte | 48 | Not recorded | 698abbb0c97a0b3e3b5de053 | | TOKENSWAP | 47 | Not recorded | 698abbb0c97a0b3e3b5de053 | | RCAD | 46 | Not recorded | 698abbb0c97a0b3e3b5de053 | | TOKENSWAP | 44 | Not recorded | 698abbb0c97a0b3e3b5de053 | | TOKENSWAP | 42 | Not recorded | 698abbb0c97a0b3e3b5de053 | | MEMFREE | 37 | Not recorded | 698abbb0c97a0b3e3b5de053 | | RCAD | 37 | Not recorded | 698abbb0c97a0b3e3b5de053 | | CP-FUSE | 37 | Not recorded | 698abbb0c97a0b3e3b5de053 | | TOKENSWAP | 37 | Not recorded | 698abbb0c97a0b3e3b5de053 | | RCAD | 36 | Not recorded | 698abbb0c97a0b3e3b5de053 | | RCAD | 36 | Not recorded | 698abbb0c97a0b3e3b5de053 | | CP-FUSE | 27 | Not recorded | 698abbb0c97a0b3e3b5de053 | | CP-FUSE | 25 | Not recorded | 698abbb0c97a0b3e3b5de053 | | CP-FUSE | 23 | Not recorded | 698abbb0c97a0b3e3b5de053 | | CP-FUSE | 20 | Not recorded | 698abbb0c97a0b3e3b5de053 | | CP-FUSE | 18 | Not recorded | 698abbb0c97a0b3e3b5de053 | | SAFE | 16 | Not recorded | 698abbb0c97a0b3e3b5de053 | | SAFE | 16 | Not recorded | 698abbb0c97a0b3e3b5de053 | | SAFE | 16 | Not recorded | 698abbb0c97a0b3e3b5de053 | | SAFE | 9 | Not recorded | 698abbb0c97a0b3e3b5de053 | | SAFE | 9 | Not recorded | 698abbb0c97a0b3e3b5de053 | | SAFE | 9 | Not recorded | 698abbb0c97a0b3e3b5de053 | ### PROXIMALByte Factuality: 56. Result ID: 699643bb5698665ec4875dad. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Safe model (ps)":"TINYCOMMA 1.8B","Risky model (pr)":"LLAMA 4 17B×16E","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### PROXIMALByte Factuality: 56. Result ID: 699643bb5698665ec4875dae. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Safe model (ps)":"COMMA 7B","Risky model (pr)":"LLAMA 4 17B×16E","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### PROXIMAL Factuality: 53. Result ID: 699643bb5698665ec4875da9. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Safe model (ps)":"TINYCOMMA 1.8B","Risky model (pr)":"LLAMA 3.1 70B","Decoding level":"Token-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### PROXIMALByte Factuality: 52. Result ID: 699643bb5698665ec4875daa. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Safe model (ps)":"COMMA 7B","Risky model (pr)":"LLAMA 3.1 70B","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### TOKENSWAP Factuality: 49. Result ID: 699643bb5698665ec4875da5. [Reporting paper]() Recorded evaluation configuration: {"Safe model (ps)":"COMMA 7B","Risky model (pr)":"LLAMA 3.1 70B","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### PROXIMALByte Factuality: 49. Result ID: 699643bb5698665ec4875dac. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Safe model (ps)":"COMMA 7B","Risky model (pr)":"QWEN 2.5 72B","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### PROXIMALByte Factuality: 48. Result ID: 699643bb5698665ec4875dab. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Safe model (ps)":"TINYCOMMA 1.8B","Risky model (pr)":"QWEN 2.5 72B","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### TOKENSWAP Factuality: 47. Result ID: 699643bb5698665ec4875da8. [Reporting paper]() Recorded evaluation configuration: {"Safe model (ps)":"TINYCOMMA 1.8B","Risky model (pr)":"LLAMA 4 17B×16E","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### RCAD Factuality: 46. Result ID: 699643bb5698665ec4875d9b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Safe model (ps)":"COMMA 7B","Risky model (pr)":"LLAMA 3.1 70B","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### TOKENSWAP Factuality: 44. Result ID: 699643bb5698665ec4875da4. [Reporting paper]() Recorded evaluation configuration: {"Safe model (ps)":"TINYCOMMA 1.8B","Risky model (pr)":"LLAMA 3.1 70B","Decoding level":"Token-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### TOKENSWAP Factuality: 42. Result ID: 699643bb5698665ec4875da7. [Reporting paper]() Recorded evaluation configuration: {"Safe model (ps)":"COMMA 7B","Risky model (pr)":"QWEN 2.5 72B","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### MEMFREE Factuality: 37. Result ID: 699643bb5698665ec4875d99. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Safe model (ps)":"TINYCOMMA 1.8B","Risky model (pr)":"LLAMA 3.1 70B","Decoding level":"Token-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### RCAD Factuality: 37. Result ID: 699643bb5698665ec4875d9a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Safe model (ps)":"TINYCOMMA 1.8B","Risky model (pr)":"LLAMA 3.1 70B","Decoding level":"Token-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### CP-FUSE Factuality: 37. Result ID: 699643bb5698665ec4875da3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Safe model (ps)":"COMMA 7B","Risky model (pr)":"LLAMA 4 17B×16E","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### TOKENSWAP Factuality: 37. Result ID: 699643bb5698665ec4875da6. [Reporting paper]() Recorded evaluation configuration: {"Safe model (ps)":"TINYCOMMA 1.8B","Risky model (pr)":"QWEN 2.5 72B","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### RCAD Factuality: 36. Result ID: 699643bb5698665ec4875d9c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Safe model (ps)":"TINYCOMMA 1.8B","Risky model (pr)":"LLAMA 4 17B×16E","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### RCAD Factuality: 36. Result ID: 699643bb5698665ec4875d9d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Safe model (ps)":"COMMA 7B","Risky model (pr)":"LLAMA 4 17B×16E","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### CP-FUSE Factuality: 27. Result ID: 699643bb5698665ec4875da2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Safe model (ps)":"TINYCOMMA 1.8B","Risky model (pr)":"LLAMA 4 17B×16E","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### CP-FUSE Factuality: 25. Result ID: 699643bb5698665ec4875da1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Safe model (ps)":"COMMA 7B","Risky model (pr)":"QWEN 2.5 72B","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### CP-FUSE Factuality: 23. Result ID: 699643bb5698665ec4875d9f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Safe model (ps)":"COMMA 7B","Risky model (pr)":"LLAMA 3.1 70B","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### CP-FUSE Factuality: 20. Result ID: 699643bb5698665ec4875d9e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Safe model (ps)":"TINYCOMMA 1.8B","Risky model (pr)":"LLAMA 3.1 70B","Decoding level":"Token-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### CP-FUSE Factuality: 18. Result ID: 699643bb5698665ec4875da0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Safe model (ps)":"TINYCOMMA 1.8B","Risky model (pr)":"QWEN 2.5 72B","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### SAFE Factuality: 16. Result ID: 699643bb5698665ec4875d94. [Reporting paper]() Recorded evaluation configuration: {"Safe model (ps)":"COMMA 7B","Risky model (pr)":"LLAMA 3.1 70B","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### SAFE Factuality: 16. Result ID: 699643bb5698665ec4875d96. [Reporting paper]() Recorded evaluation configuration: {"Safe model (ps)":"COMMA 7B","Risky model (pr)":"QWEN 2.5 72B","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### SAFE Factuality: 16. Result ID: 699643bb5698665ec4875d98. [Reporting paper]() Recorded evaluation configuration: {"Safe model (ps)":"COMMA 7B","Risky model (pr)":"LLAMA 4 17B×16E","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### SAFE Factuality: 9. Result ID: 699643bb5698665ec4875d93. [Reporting paper]() Recorded evaluation configuration: {"Safe model (ps)":"TINYCOMMA 1.8B","Risky model (pr)":"LLAMA 3.1 70B","Decoding level":"Token-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### SAFE Factuality: 9. Result ID: 699643bb5698665ec4875d95. [Reporting paper]() Recorded evaluation configuration: {"Safe model (ps)":"TINYCOMMA 1.8B","Risky model (pr)":"QWEN 2.5 72B","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. ### SAFE Factuality: 9. Result ID: 699643bb5698665ec4875d97. [Reporting paper]() Recorded evaluation configuration: {"Safe model (ps)":"TINYCOMMA 1.8B","Risky model (pr)":"LLAMA 4 17B×16E","Decoding level":"Byte-level","NCR threshold":"\>= 75%"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.