# Peer Review Feedback Generation on ICLR papers Source: [Peer Review Feedback Generation on ICLR papers]() Collection index: [llms.txt]() Data updated: 2026-04-15T07:42:16.041Z Evaluation of model performance on generating peer review feedback for ICLR papers, measured by combined success rates and validity rates. 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 | Peer Review Feedback Generation | | Dataset (including any stated split) | ICLR papers | | Metric | Combined Success Rate | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 6 of 6 | | 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 - [Combined Success Rate](): Higher is better; unit: PERCENT. - [Combined Success Confidence Interval (CI)](): Lower is better; unit: PERCENT. - [Validity Only Rate](): Higher is better; unit: PERCENT. - [Validity Only Confidence Interval (CI)](): Lower is better; unit: PERCENT. - [Author Action Only Rate](): Higher is better; unit: PERCENT. - [Author Action Only Confidence Interval (CI)](): Lower is better; unit: PERCENT. ## Recorded results | Method | Combined Success Rate | Result date | Report paper | | --- | --- | --- | --- | | GPT-5.2 | 45.8 | Not recorded | 69df1b967d0172140ec7b901 | | Gemini-3-flash | 37.9 | Not recorded | 69df1b967d0172140ec7b901 | | GOODPOINT-DPO | 14.7 | Not recorded | 69df1b967d0172140ec7b901 | | GOODPOINT-SFT | 9.2 | Not recorded | 69df1b967d0172140ec7b901 | | Qwen3-8b (Base) | 8 | Not recorded | 69df1b967d0172140ec7b901 | | Llama3.1-8b-Instruct | 1.8 | Not recorded | 69df1b967d0172140ec7b901 | ### GPT-5.2 Combined Success Rate: 45.8. Result ID: 69df3fea47f34ce924c7b62d. [Reporting paper]() Recorded evaluation configuration: {"Bootstrap iterations (B)":"1000","Feedback per paper":"5"}. This record does not assert independent reproduction. ### Gemini-3-flash Combined Success Rate: 37.9. Result ID: 69df3fea47f34ce924c7b62c. [Reporting paper]() Recorded evaluation configuration: {"Bootstrap iterations (B)":"1000","Feedback per paper":"5"}. This record does not assert independent reproduction. ### GOODPOINT-DPO Combined Success Rate: 14.7. Result ID: 69df3fea47f34ce924c7b630. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Protocol":"DPO","Bootstrap iterations (B)":"1000","Feedback per paper":"5"}. This record does not assert independent reproduction. ### GOODPOINT-SFT Combined Success Rate: 9.2. Result ID: 69df3fea47f34ce924c7b631. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Training Protocol":"SFT","Bootstrap iterations (B)":"1000","Feedback per paper":"5"}. This record does not assert independent reproduction. ### Qwen3-8b (Base) Combined Success Rate: 8. Result ID: 69df3fea47f34ce924c7b62f. [Reporting paper]() Recorded evaluation configuration: {"Model Scale":"8b","Training Protocol":"Base","Bootstrap iterations (B)":"1000","Feedback per paper":"5"}. This record does not assert independent reproduction. ### Llama3.1-8b-Instruct Combined Success Rate: 1.8. Result ID: 69df3fea47f34ce924c7b62e. [Reporting paper]() Recorded evaluation configuration: {"Model Scale":"8b","Training Protocol":"Instruct","Bootstrap iterations (B)":"1000","Feedback per paper":"5"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.