Out-of-distribution detection on OOD Detection Source LLM: Claude-3.5-Haiku (test)
0.977XSUMREADER
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| READERevaluation_config=per-domain2026.05 | 0.977 | 0.923 | 0.873 | 0.963 | 0.934 | |
| RADARevaluation_config=per-domain2026.05 | 0.857 | 0.72 | 0.777 | 0.847 | 0.8 | |
| ADADETECTGPTtraining_variant=fixed-training, surrogate_llm=Gemma, evaluation_config=per-domain2026.05 | 0.837 | 0.87 | 0.923 | 0.9 | 0.882 | |
| BINOCULARSevaluation_config=per-domain2026.05 | 0.833 | 0.88 | 0.92 | 0.903 | 0.884 | |
| FAST-DETECTGPTsurrogate_llm=Gemma, evaluation_config=per-domain2026.05 | 0.833 | 0.873 | 0.927 | 0.897 | 0.882 | |
| IMBDtraining_variant=fixed-training, evaluation_config=per-domain2026.05 | 0.823 | 0.917 | 0.913 | 0.873 | 0.882 | |
| BISCOPEtraining_variant=fixed-training, evaluation_config=per-domain2026.05 | 0.78 | 0.787 | 0.707 | 0.817 | 0.772 | |
| BISCOPEtraining_variant=target-adapted, evaluation_config=per-domain2026.05 | 0.773 | 0.76 | 0.737 | 0.77 | 0.76 | |
| IMBDtraining_variant=target-adapted, evaluation_config=per-domain2026.05 | 0.72 | 0.743 | 0.75 | 0.787 | 0.75 | |
| DETECTGPTsurrogate_llm=Gemma, evaluation_config=per-domain2026.05 | 0.663 | 0.74 | 0.703 | 0.807 | 0.728 | |
| NPRsurrogate_llm=Gemma, evaluation_config=per-domain2026.05 | 0.613 | 0.73 | 0.58 | 0.803 | 0.682 | |
| ENTROPYsurrogate_llm=Gemma, evaluation_config=per-domain2026.05 | 0.593 | 0.59 | 0.577 | 0.63 | 0.598 | |
| DNAGPTsurrogate_llm=Gemma, evaluation_config=per-domain2026.05 | 0.573 | 0.64 | 0.773 | 0.807 | 0.698 | |
| ROBERTALARGEtraining_variant=target-adapted, evaluation_config=per-domain2026.05 | 0.567 | 0.593 | 0.543 | 0.6 | 0.576 | |
| ROBERTABASEtraining_variant=target-adapted, evaluation_config=per-domain2026.05 | 0.55 | 0.603 | 0.56 | 0.587 | 0.575 | |
| LIKELIHOODsurrogate_llm=Gemma, evaluation_config=per-domain2026.05 | 0.537 | 0.637 | 0.7 | 0.66 | 0.633 | |
| LRRsurrogate_llm=Gemma, evaluation_config=per-domain2026.05 | 0.537 | 0.567 | 0.557 | 0.623 | 0.571 | |
| LOGRANKsurrogate_llm=Gemma, evaluation_config=per-domain2026.05 | 0.53 | 0.627 | 0.683 | 0.657 | 0.624 | |
| ADADETECTGPTtraining_variant=target-adapted, surrogate_llm=Gemma, evaluation_config=per-domain2026.05 | 0.53 | 0.6 | 0.537 | 0.597 | 0.566 |