UKPLab/ProReviewer-8B
Text Generation • 8B • Updated
Datasets:
paper_id string | arxiv_id string | title string | markdown dict | reviews list | scores dict | metadata dict | meta_review dict | decision dict |
|---|---|---|---|---|---|---|---|---|
00SnKBGTsz | 2410.06215v1 | DataEnvGym: Data Generation Agents in Teacher Environments with Student Feedback | {
"content": "## Abstract\n\nAbstract The process of creating training data to teach models is currently driven by humans, who manually analyze model weaknesses and plan how to create data that improves a student model.\nRecent approaches using large language models (LLMs) as annotators reduce human annotation effort... | [
{
"id": "r8ZflFk3T7",
"initial_rating": 8,
"confidence": 4,
"soundness": 4,
"contribution": 4,
"presentation": 3,
"summary": "This paper introduces Gym environments for data synthesis, framing the problem as sequential decision-making. In these environments, actions correspond to data-ge... | {
"rating": "5;6;6;8",
"rating_avg": 6.25,
"confidence": "4;3;4;4",
"confidence_avg": 3.75,
"soundness": "2;2;4;3",
"soundness_avg": 2.75,
"contribution": "3;3;4;3",
"contribution_avg": 3.25,
"presentation": "3;3;2;2",
"presentation_avg": 2.5
} | {
"primary_area": "",
"track": "main",
"venue": "ICLR 2025 Spotlight",
"venueid": "ICLR.cc/2025/Conference",
"processed_at": "2026-01-14T22:16:00.543857"
} | {
"id": "zpboemkkjR",
"metareview": "The paper frames the problem of automatic data generation (to improve a ML model) as a sequential decision making task, and provides Gym environments as well as LLM-based agents that are effective for them. The resulting datasets are shown to be effective for ML models in math r... | {
"decision": "Accept (Spotlight)"
} |
00ezkB2iZf | 2406.06573v2 | MedFuzz: Exploring the Robustness of Large Language Models in Medical Question Answering | {
"content": "## Abstract\n\nAbstract Large language models (LLM) have achieved impressive performance on medical question-answering benchmarks.\nHowever, high benchmark accuracy does not imply that the performance generalizes to real-world clinical settings.\nMedical question-answering benchmarks rely on assumptions... | [
{
"id": "TeO25XUwES",
"initial_rating": 3,
"confidence": 4,
"soundness": 2,
"contribution": 2,
"presentation": 2,
"summary": "This paper investigates the robustness of large language models in handling medical QA tasks by introducing a new evaluation method, MedFuzz. For each multiple-ch... | {
"rating": "3;3;5;6",
"rating_avg": 4.25,
"confidence": "4;4;5;3",
"confidence_avg": 4,
"soundness": "3;2;3;3",
"soundness_avg": 2.75,
"contribution": "2;2;4;3",
"contribution_avg": 2.75,
"presentation": "3;2;3;3",
"presentation_avg": 2.75
} | {
"primary_area": "",
"track": "main",
"venue": "Submitted to ICLR 2025",
"venueid": "ICLR.cc/2025/Conference/Rejected_Submission",
"processed_at": "2026-01-14T22:16:00.544972"
} | {
"id": "Se1GK2iVy4",
"metareview": "In this paper, the authors propose MedFuzz, an LLM-based technique to provide challenging medical questions. The technique allows to test medical LLMs at scale on a multichoice QA dataset, by modifying factors deemed as 'irrelevant' to the final diagnosis.\n\nWhile all reviewers... | {
"decision": "Reject"
} |
02Od16GFRW | 2410.01452v1 | Ensembles provably learn equivariance through data augmentation | {"content":"## Abstract\n\nAbstract Recently, it was proved that group equivariance emerges in ensem(...TRUNCATED) | [{"id":"HKJJNQ1JKw","initial_rating":3,"confidence":4,"soundness":3,"contribution":2,"presentation":(...TRUNCATED) | {"rating":"3;6;6","rating_avg":5.0,"confidence":"4;3;3","confidence_avg":3.3333333333333335,"soundne(...TRUNCATED) | {"primary_area":"","track":"main","venue":"Submitted to ICLR 2025","venueid":"ICLR.cc/2025/Conferenc(...TRUNCATED) | {"id":"I3eiVnCVfC","metareview":"This paper studies how equivariance emerges in ensembles of neural (...TRUNCATED) | {
"decision": "Reject"
} |
02haSpO453 | 2409.04429v2 | VILA-U: a Unified Foundation Model Integrating Visual Understanding and Generation | {"content":"## Abstract\n\nAbstract VILA-U is a U nified foundation model that integrates V ideo, I (...TRUNCATED) | [{"id":"cGas6kZlaM","initial_rating":6,"confidence":4,"soundness":4,"contribution":3,"presentation":(...TRUNCATED) | {"rating":"3;5;5;6","rating_avg":4.75,"confidence":"4;5;3;4","confidence_avg":4.0,"soundness":"3;2;2(...TRUNCATED) | {"primary_area":"","track":"main","venue":"ICLR 2025 Poster","venueid":"ICLR.cc/2025/Conference","pr(...TRUNCATED) | {"id":"gg4i7pnNPQ","metareview":"VILA-U presents a unified foundation model that integrates video, i(...TRUNCATED) | {
"decision": "Accept (Poster)"
} |
02kZwCo0C3 | 2406.15567v1 | SAIL: Self-improving Efficient Online Alignment of Large Language Models | {"content":"## Abstract\n\nAbstract Reinforcement Learning from Human Feedback (RLHF) is a key metho(...TRUNCATED) | [{"id":"BU6la6v4Ci","initial_rating":6,"confidence":4,"soundness":3,"contribution":4,"presentation":(...TRUNCATED) | {"rating":"3;6;6;8","rating_avg":5.75,"confidence":"4;3;4;4","confidence_avg":3.75,"soundness":"3;3;(...TRUNCATED) | {"primary_area":"","track":"main","venue":"Submitted to ICLR 2025","venueid":"ICLR.cc/2025/Conferenc(...TRUNCATED) | {"id":"E3XYUMBr1C","metareview":"The paper introduces SAIL, a self-improving online RLHF approach fo(...TRUNCATED) | {
"decision": "Reject"
} |
03EkqSCKuO | 2405.17163v1 | Port-Hamiltonian Architectural Bias for Long-Range Propagation in Deep Graph Networks | {"content":"## Abstract\n\nAbstract The dynamics of information diffusion within graphs is a critica(...TRUNCATED) | [{"id":"d6JJf0KjwN","initial_rating":6,"confidence":3,"soundness":4,"contribution":2,"presentation":(...TRUNCATED) | {"rating":"5;6;8","rating_avg":6.333333333333333,"confidence":"2;3;3","confidence_avg":2.66666666666(...TRUNCATED) | {"primary_area":"","track":"main","venue":"ICLR 2025 Poster","venueid":"ICLR.cc/2025/Conference","pr(...TRUNCATED) | {"id":"H9PCwNoHeA","metareview":"**(a) Scientific Claims and Findings:**\nThe paper introduces a nov(...TRUNCATED) | {
"decision": "Accept (Poster)"
} |
03u7pbpyeN | 2409.17972v2 | "BEATS: Optimizing LLM Mathematical Capabilities with BackVerify and Adaptive Disambiguate based Eff(...TRUNCATED) | {"content":"## Abstract\n\nAbstract Large Language Models (LLMs) have exhibited exceptional performa(...TRUNCATED) | [{"id":"0Md90Alr6g","initial_rating":3,"confidence":3,"soundness":3,"contribution":2,"presentation":(...TRUNCATED) | {"rating":"3;3;5;6","rating_avg":4.25,"confidence":"5;3;4;4","confidence_avg":4.0,"soundness":"3;3;4(...TRUNCATED) | {"primary_area":"","track":"main","venue":"ICLR 2025 Conference Withdrawn Submission","venueid":"ICL(...TRUNCATED) | {
"id": "",
"metareview": "",
"additional_comments": ""
} | {
"decision": ""
} |
06GH83hDIv | 2410.01871v1 | Auction-Based Regulation for Artificial Intelligence | {"content":"## Abstract\n\nAbstract In an era of “moving fast and breaking things” , regulators (...TRUNCATED) | [{"id":"54x9fXc8nh","initial_rating":6,"confidence":3,"soundness":3,"contribution":3,"presentation":(...TRUNCATED) | {"rating":"3;5;5;8","rating_avg":5.25,"confidence":"3;4;3;3","confidence_avg":3.25,"soundness":"2;2;(...TRUNCATED) | {"primary_area":"","track":"main","venue":"Submitted to ICLR 2025","venueid":"ICLR.cc/2025/Conferenc(...TRUNCATED) | {"id":"J6U2HHgjTm","metareview":"This paper looks at the problem of regulating AI models, specifical(...TRUNCATED) | {
"decision": "Reject"
} |
07yvxWDSla | 2409.07431v2 | Synthetic continued pretraining | {"content":"## Abstract\n\nAbstract Pretraining on large-scale, unstructured internet text enables l(...TRUNCATED) | [{"id":"5lDBbDNEH3","initial_rating":8,"confidence":4,"soundness":4,"contribution":3,"presentation":(...TRUNCATED) | {"rating":"6;6;8;8","rating_avg":7.0,"confidence":"3;4;3;4","confidence_avg":3.5,"soundness":"3;2;4;(...TRUNCATED) | {"primary_area":"","track":"main","venue":"ICLR 2025 Oral","venueid":"ICLR.cc/2025/Conference","proc(...TRUNCATED) | {"id":"eEZ1z1zkVr","metareview":"The paper proposes a data-synthesizing method for continued pretrai(...TRUNCATED) | {
"decision": "Accept (Oral)"
} |
09FiNmvNMw | 2410.08047v1 | "Divide and Translate: Compositional First-Order Logic Translation and Verification for Complex Logi(...TRUNCATED) | {"content":"## Abstract\n\nAbstract Complex logical reasoning tasks require a long sequence of reaso(...TRUNCATED) | [{"id":"o6kdM9TOlc","initial_rating":5,"confidence":4,"soundness":3,"contribution":2,"presentation":(...TRUNCATED) | {"rating":"5;5;5","rating_avg":5.0,"confidence":"4;3;4","confidence_avg":3.6666666666666665,"soundne(...TRUNCATED) | {"primary_area":"","track":"main","venue":"ICLR 2025 Poster","venueid":"ICLR.cc/2025/Conference","pr(...TRUNCATED) | {"id":"yQMwgvFDdg","metareview":"The reviewers generally saw the merits of the proposal. There was g(...TRUNCATED) | {
"decision": "Accept (Poster)"
} |
End of preview. Expand in Data Studio
ProReviewer Dataset
A version-matched dataset of ICLR paper submissions paired with their initial peer reviews and scores. Each sample contains the initial submission of a paper matched with the reviewers' initial ratings (before the discussion phase), enabling research on automated scientific peer review.
Dataset Description
This dataset supports training and evaluating automated peer review systems. All papers and reviews are version-matched: the paper content corresponds to the initial submission, and the review scores reflect the initial ratings assigned before any author-reviewer discussion.
Splits
| Split | Source | Samples | Description |
|---|---|---|---|
| train | ICLR 2025 | 4,011 | Training set for RL/SFT |
| test | ICLR 2026 | 1,000 | Held-out evaluation set |
Features
| Field | Type | Description |
|---|---|---|
| paper_id | string | OpenReview paper ID |
| arxiv_id | string | arXiv identifier |
| title | string | Paper title |
| markdown | dict | Paper content (initial submission) in markdown with metadata (authors, char/line counts) |
| reviews | list | List of peer reviews (see below) |
| scores | dict | Aggregated initial scores across reviewers (rating_avg, confidence_avg, etc.) |
| metadata | dict | Submission metadata (primary_area, track, venue) |
| meta_review | dict | Area chair meta-review and additional comments |
| decision | dict | Final acceptance decision (e.g., Accept (Poster), Reject) |
Each review contains:
| Field | Type | Description |
|---|---|---|
| id | string | Reviewer ID |
| initial_rating | int | Rating before discussion (1-10) |
| confidence | int | Reviewer confidence score |
| soundness | int | Technical soundness score |
| contribution | int | Contribution significance score |
| presentation | int | Presentation quality score |
| summary | string | Paper summary by reviewer |
| strengths | string | Identified strengths |
| weaknesses | string | Identified weaknesses |
| questions | string | Questions for the authors |
Usage
from datasets import load_dataset
ds = load_dataset("UKPLab/ProReviewer-Dataset")
# Access a training sample
paper = ds["train"][0]
print(paper["title"])
print(f"Avg initial rating: {paper['scores']['rating_avg']}")
print(f"Decision: {paper['decision']['decision']}")
print(f"Number of reviews: {len(paper['reviews'])}")
# Read paper content (initial submission)
content = paper["markdown"]["content"]
# Access individual reviews
for review in paper["reviews"]:
print(f"Initial rating: {review['initial_rating']}, Confidence: {review['confidence']}")
Version Matching
All data is version-matched to ensure consistency:
- Paper content: Corresponds to the initial submission (before any revisions in response to reviews)
- Review scores: Reflect the initial ratings assigned by reviewers (before the discussion/rebuttal phase)
This version matching is important for training review agents, as it ensures the model learns to evaluate papers as they were first submitted, with scores that reflect first impressions rather than post-discussion adjustments.
Source
Papers and reviews were collected from OpenReview for the ICLR 2025 and ICLR 2026 venues.
Associated Resources
- Paper: From Passive Generation to Investigation: A Proactive Scientific Peer Review Agent
- Code: UKPLab/arxiv2026-ProReviewer
Citation
@article{fang2026passive,
title={From Passive Generation to Investigation: A Proactive Scientific Peer Review Agent},
author={Fang, Haishuo and Feng, Yue and Gurevych, Iryna},
journal={arXiv preprint arXiv:2606.13349},
year={2026}
}
License
This dataset is released under the MIT License.
- Downloads last month
- -
Models trained or fine-tuned on UKPLab/ProReviewer-Dataset