# Boundary Detection on RoFT-chatgpt GPT-3.5-turbo generated (test) Source: [Boundary Detection on RoFT-chatgpt GPT-3.5-turbo generated (test)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:30:09.065Z Evaluation of Boundary Detection performance using results generated by GPT-3.5-turbo for the RoFT benchmark suite (test split). Metrics include Acc, SoftAcc1, and MSE. 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 | Boundary Detection | | Dataset (including any stated split) | RoFT-chatgpt GPT-3.5-turbo generated (test) | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 34 of 34 | | 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 - [Accuracy](): Higher is better; unit: PERCENT. - [Soft Accuracy 1](): Higher is better; unit: PERCENT. - [Mean Squared Error](): Lower is better; unit: SCALAR. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | GigaCheck (DN-DAB-DETR) | 67.65 | Not recorded | 68fa75991e209131cea68dbf | | ROBERTa + SEP | 54.61 | Not recorded | 68fa74c21e209131cea63965 | | RoBERTa + SEP | 54.61 | Not recorded | 68fa75991e209131cea68dbf | | Phi-1.5 Perpl. + GB classifier | 50.8 | Not recorded | 68fa74c21e209131cea63965 | | Phi-2 Perpl. + GB classifier | 48.4 | Not recorded | 68fa74c21e209131cea63965 | | Phi-1.5 Perpl. + LR classifier | 47.3 | Not recorded | 68fa74c21e209131cea63965 | | Phi-1.5 | 47.3 | Not recorded | 68fa75991e209131cea68dbf | | Phi-2 Perpl. + LR classifier | 44.4 | Not recorded | 68fa74c21e209131cea63965 | | LLaMA-2-7B Perpl. + GB classifier | 41.5 | Not recorded | 68fa74c21e209131cea63965 | | ROBERTa | 39.01 | Not recorded | 68fa74c21e209131cea63965 | | RoBERTa | 39.01 | Not recorded | 68fa75991e209131cea68dbf | | Phi-1 Perpl. + GB classifier | 36.5 | Not recorded | 68fa74c21e209131cea63965 | | GPT-2 Perpl. + GB classifier | 34.94 | Not recorded | 68fa74c21e209131cea63965 | | GPT-2 Perpl. + LR classifier | 33.5 | Not recorded | 68fa74c21e209131cea63965 | | Phi-1.5 Perpl. + GB regressor | 32 | Not recorded | 68fa74c21e209131cea63965 | | Phi-1.5 | 32 | Not recorded | 68fa75991e209131cea68dbf | | Phi-1 Perpl. + LR classifier | 31 | Not recorded | 68fa74c21e209131cea63965 | | Phi-2 Perpl. + GB regressor | 30.8 | Not recorded | 68fa74c21e209131cea63965 | | LLaMA-2-7B Perpl. + LR classifier | 29.7 | Not recorded | 68fa74c21e209131cea63965 | | Length + GB | 25.72 | Not recorded | 68fa74c21e209131cea63965 | | LLaMA-2-7B Perpl. + GB regressor | 23.6 | Not recorded | 68fa74c21e209131cea63965 | | DetectGPT + GB classifier | 21.69 | Not recorded | 68fa74c21e209131cea63965 | | DetectGPT | 21.69 | Not recorded | 68fa75991e209131cea68dbf | | TLE + TS Binary | 20.02 | Not recorded | 68fa74c21e209131cea63965 | | TLE + TS Binary | 20.02 | Not recorded | 68fa75991e209131cea68dbf | | GPT-2 Perpl. + GB regressor | 19.74 | Not recorded | 68fa74c21e209131cea63965 | | Phi-1 Perpl. + GB regressor | 17.4 | Not recorded | 68fa74c21e209131cea63965 | | PHD + TS ML | 17.29 | Not recorded | 68fa74c21e209131cea63965 | | PHD + TS ML | 17.29 | Not recorded | 68fa75991e209131cea68dbf | | DetectGPT + LR classifier | 15.35 | Not recorded | 68fa74c21e209131cea63965 | | DetectGPT | 15.35 | Not recorded | 68fa75991e209131cea68dbf | | Majority class | 13.83 | Not recorded | 68fa74c21e209131cea63965 | | gpt-4 | 9.1 | Not recorded | 68fa74c21e209131cea63965 | | gpt-3.5-turbo | 7.1 | Not recorded | 68fa74c21e209131cea63965 | ### GigaCheck (DN-DAB-DETR) Accuracy: 67.65. Result ID: 6993f17c3a6c31484919aa7b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### ROBERTa + SEP Accuracy: 54.61. Result ID: 699359d33a6c3148491552da. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### RoBERTa + SEP Accuracy: 54.61. Result ID: 6993f17c3a6c31484919aa79. [Reporting paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### Phi-1.5 Perpl. + GB classifier Accuracy: 50.8. Result ID: 699359d33a6c3148491552e4. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"category":"Based on Perplexity"}. This record does not assert independent reproduction. ### Phi-2 Perpl. + GB classifier Accuracy: 48.4. Result ID: 699359d33a6c3148491552e7. [Reporting paper]() Recorded evaluation configuration: {"category":"Based on Perplexity"}. This record does not assert independent reproduction. ### Phi-1.5 Perpl. + LR classifier Accuracy: 47.3. Result ID: 699359d33a6c3148491552e6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"category":"Based on Perplexity"}. This record does not assert independent reproduction. ### Phi-1.5 Accuracy: 47.3. Result ID: 6993f17c3a6c31484919aa7f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"classifier":"LR classifier","type":"Perplexity-based"}. This record does not assert independent reproduction. ### Phi-2 Perpl. + LR classifier Accuracy: 44.4. Result ID: 699359d33a6c3148491552e9. [Reporting paper]() Recorded evaluation configuration: {"category":"Based on Perplexity"}. This record does not assert independent reproduction. ### LLaMA-2-7B Perpl. + GB classifier Accuracy: 41.5. Result ID: 699359d33a6c3148491552ea. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"category":"Based on Perplexity"}. This record does not assert independent reproduction. ### ROBERTa Accuracy: 39.01. Result ID: 699359d33a6c3148491552db. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### RoBERTa Accuracy: 39.01. Result ID: 6993f17c3a6c31484919aa7a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### Phi-1 Perpl. + GB classifier Accuracy: 36.5. Result ID: 699359d33a6c3148491552e1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"category":"Based on Perplexity"}. This record does not assert independent reproduction. ### GPT-2 Perpl. + GB classifier Accuracy: 34.94. Result ID: 699359d33a6c3148491552de. [Reporting paper]() Recorded evaluation configuration: {"category":"Based on Perplexity"}. This record does not assert independent reproduction. ### GPT-2 Perpl. + LR classifier Accuracy: 33.5. Result ID: 699359d33a6c3148491552e0. [Reporting paper]() Recorded evaluation configuration: {"category":"Based on Perplexity"}. This record does not assert independent reproduction. ### Phi-1.5 Perpl. + GB regressor Accuracy: 32. Result ID: 699359d33a6c3148491552e5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"category":"Based on Perplexity"}. This record does not assert independent reproduction. ### Phi-1.5 Accuracy: 32. Result ID: 6993f17c3a6c31484919aa7e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"regressor":"GB regressor","type":"Perplexity-based"}. This record does not assert independent reproduction. ### Phi-1 Perpl. + LR classifier Accuracy: 31. Result ID: 699359d33a6c3148491552e3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"category":"Based on Perplexity"}. This record does not assert independent reproduction. ### Phi-2 Perpl. + GB regressor Accuracy: 30.8. Result ID: 699359d33a6c3148491552e8. [Reporting paper]() Recorded evaluation configuration: {"category":"Based on Perplexity"}. This record does not assert independent reproduction. ### LLaMA-2-7B Perpl. + LR classifier Accuracy: 29.7. Result ID: 699359d33a6c3148491552ec. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"category":"Based on Perplexity"}. This record does not assert independent reproduction. ### Length + GB Accuracy: 25.72. Result ID: 699359d33a6c3148491552f1. [Reporting paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### LLaMA-2-7B Perpl. + GB regressor Accuracy: 23.6. Result ID: 699359d33a6c3148491552eb. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"category":"Based on Perplexity"}. This record does not assert independent reproduction. ### DetectGPT + GB classifier Accuracy: 21.69. Result ID: 699359d33a6c3148491552dc. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"category":"Based on Perplexity"}. This record does not assert independent reproduction. ### DetectGPT Accuracy: 21.69. Result ID: 6993f17c3a6c31484919aa7c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"classifier":"GB classifier","type":"Perplexity-based"}. This record does not assert independent reproduction. ### TLE + TS Binary Accuracy: 20.02. Result ID: 699359d33a6c3148491552ee. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"category":"Based on TDA"}. This record does not assert independent reproduction. ### TLE + TS Binary Accuracy: 20.02. Result ID: 6993f17c3a6c31484919aa81. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"type":"TDA-based"}. This record does not assert independent reproduction. ### GPT-2 Perpl. + GB regressor Accuracy: 19.74. Result ID: 699359d33a6c3148491552df. [Reporting paper]() Recorded evaluation configuration: {"category":"Based on Perplexity"}. This record does not assert independent reproduction. ### Phi-1 Perpl. + GB regressor Accuracy: 17.4. Result ID: 699359d33a6c3148491552e2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"category":"Based on Perplexity"}. This record does not assert independent reproduction. ### PHD + TS ML Accuracy: 17.29. Result ID: 699359d33a6c3148491552ed. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"category":"Based on TDA"}. This record does not assert independent reproduction. ### PHD + TS ML Accuracy: 17.29. Result ID: 6993f17c3a6c31484919aa80. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"type":"TDA-based"}. This record does not assert independent reproduction. ### DetectGPT + LR classifier Accuracy: 15.35. Result ID: 699359d33a6c3148491552dd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"category":"Based on Perplexity"}. This record does not assert independent reproduction. ### DetectGPT Accuracy: 15.35. Result ID: 6993f17c3a6c31484919aa7d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"classifier":"LR classifier","type":"Perplexity-based"}. This record does not assert independent reproduction. ### Majority class Accuracy: 13.83. Result ID: 699359d33a6c3148491552f2. [Reporting paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### gpt-4 Accuracy: 9.1. Result ID: 699359d33a6c3148491552f0. [Reporting paper]() Recorded evaluation configuration: {"category":"Zero-shot"}. This record does not assert independent reproduction. ### gpt-3.5-turbo Accuracy: 7.1. Result ID: 699359d33a6c3148491552ef. [Reporting paper]() Recorded evaluation configuration: {"category":"Zero-shot"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.