Datasets:
File size: 12,994 Bytes
88f4cc0 05459f4 4cc847d 05459f4 4cc847d 05459f4 4cc847d 05459f4 4cc847d 88f4cc0 4cc847d 88f4cc0 05459f4 7d1b399 05459f4 628c2d9 05459f4 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 | ---
license: cc-by-4.0
task_categories:
- text-generation
- text-classification
- summarization
- question-answering
language:
- en
tags:
- peer-review
- openreview
- scientific-papers
- nlp
- benchmarking
- meta-science
pretty_name: ReviewBench — A Multi-Conference Peer-Review Corpus
size_categories:
- 10K<n<100K
configs:
- config_name: default
data_files:
- split: neurips
path: data/neurips-*
- split: iclr
path: data/iclr-*
- split: icml
path: data/icml-*
- split: tmlr
path: data/tmlr-*
- split: emnlp
path: data/emnlp-*
- split: corl
path: data/corl-*
- split: colm
path: data/colm-*
dataset_info:
features:
- name: forum_id
dtype: string
- name: conference
dtype: string
- name: year
dtype: int32
- name: track
dtype: string
- name: venue_id
dtype: string
- name: paper_number
dtype: int32
- name: title
dtype: string
- name: abstract
dtype: string
- name: authors
list: string
- name: keywords
list: string
- name: tldr
dtype: string
- name: primary_area
dtype: string
- name: venue
dtype: string
- name: decision
dtype: string
- name: decision_comment
dtype: string
- name: author_rebuttal
dtype: string
- name: num_reviews
dtype: int32
- name: reviews_json
dtype: string
- name: markdown
dtype: string
- name: markdown_chars
dtype: int64
splits:
- name: colm
num_bytes: 69559593
num_examples: 717
- name: corl
num_bytes: 55415874
num_examples: 813
- name: emnlp
num_bytes: 175864122
num_examples: 2009
- name: iclr
num_bytes: 2043218314
num_examples: 22532
- name: icml
num_bytes: 387957515
num_examples: 3257
- name: neurips
num_bytes: 2023212137
num_examples: 18453
- name: tmlr
num_bytes: 400634995
num_examples: 3748
download_size: 2728439622
dataset_size: 5155862550
---
# ReviewBench
A large, multi-conference corpus of **peer-reviewed papers + their reviews + author rebuttals + acceptance decisions**, harvested from [OpenReview](https://openreview.net) and aligned with **OCR'd full-text markdown** of every paper.
- **51,529** papers
- **196,099** reviews
- **558,785** OCR'd PDF pages (markdown inlined per row)
- **7 conferences**, **22 venue/year combinations**, **2020 – 2026**
```python
from datasets import load_dataset
ds = load_dataset("/reviewbench")
print(ds)
# DatasetDict({
# neurips: Dataset(num_rows=...)
# iclr: Dataset(num_rows=...)
# icml: Dataset(num_rows=...)
# tmlr: Dataset(num_rows=...)
# emnlp: Dataset(num_rows=...)
# corl: Dataset(num_rows=...)
# colm: Dataset(num_rows=...)
# })
```
## Coverage
One **split per conference family**. Within a split, filter by `year`, `venue_id`, or `track` for slicing.
| Split | Venues / years | Tracks | ≈ Papers |
|---------|--------------------------------------------------------------------------------------|-----------------------------------------|----------|
| neurips | 2021, 2022, 2023, 2023 D&B, 2024, 2025 | main + Datasets & Benchmarks (2023) | ~18,400 |
| iclr | 2020, 2021, 2022, 2023, 2024, 2025, 2026 | main | ~22,500 |
| icml | 2025 | main | ~3,300 |
| tmlr | rolling (all accepted papers as of April 2026) | main | ~3,750 |
| emnlp | 2023 | main + Findings | ~2,000 |
| corl | 2021, 2022, 2023, 2024 | main | ~820 |
| colm | 2024, 2025 | main | ~720 |
NeurIPS 2020 and earlier used CMT and have no public OpenReview reviews.
## Schema
Each row is one paper.
| Column | Type | Notes |
|-------------------|-----------------|-------------------------------------------------------------------------|
| `forum_id` | `string` | OpenReview forum ID — primary key |
| `conference` | `string` | `neurips` / `iclr` / `icml` / `tmlr` / `emnlp` / `corl` / `colm` |
| `year` | `int32` | Conference year |
| `track` | `string` | `main` / `datasets_and_benchmarks` / `findings` / etc. |
| `venue_id` | `string` | OpenReview venue ID, e.g. `NeurIPS.cc/2024/Conference` |
| `paper_number` | `int32` | Submission number (nullable) |
| `title` | `string` | |
| `abstract` | `string` | |
| `authors` | `list<string>` | |
| `keywords` | `list<string>` | |
| `tldr` | `string` | |
| `primary_area` | `string` | |
| `venue` | `string` | Final venue string, e.g. `"NeurIPS 2024 poster"` |
| `decision` | `string` | `Accept (poster)`, `Accept (oral)`, `Reject`, etc. |
| `decision_comment`| `string` | Area-chair meta-review |
| `author_rebuttal` | `string` | General rebuttal (≤2024); empty when per-reviewer rebuttals are used |
| `num_reviews` | `int32` | Convenience count |
| `reviews_json` | `string` | All reviews as a JSON string — see schema below |
| `markdown` | `string` | OCR'd full text of the PDF (see *OCR* below); `""` if PDF unavailable |
| `markdown_chars` | `int64` | `len(markdown)` for fast filtering |
### Decoding `reviews_json`
`reviews_json` is `json.dumps(list[dict])`. Decode with:
```python
import json
df = ds["neurips"].to_pandas()
df["reviews"] = df["reviews_json"].map(json.loads)
print(df.iloc[0]["reviews"][0].keys())
```
Each review dict is a **union over all forms used by all venues across all years**, with absent fields as empty strings / `None`. The most reliably populated fields are:
| Field | Where populated |
|-------------------|---------------------------------------------------------------|
| `review_id`, `reviewer`, `rating`, `confidence`, `rebuttal` | All venues |
| `summary`, `questions`, `limitations`, `strengths`, `weaknesses` | NeurIPS 2022–24, ICLR, ICML, CoRL, COLM |
| `soundness`, `presentation`, `contribution` | NeurIPS 2022–24, ICLR, ICML |
| `quality`, `clarity`, `significance`, `originality` | NeurIPS 2025, TMLR |
| `strengths_and_weaknesses` | NeurIPS 2025 (merged form) |
| `was_revised`, `final_justification` | NeurIPS 2025 (in-place review revisions) |
| `claims_and_evidence`, `theoretical_claims`, `experimental_designs_or_analyses`, `relation_to_broader_scientific_literature`, `essential_references_not_discussed` | TMLR |
| `paper_topic_and_main_contributions`, `reasons_to_accept`, `reasons_to_reject`, `excitement`, `reproducibility`, `ethical_concerns` | EMNLP 2023 |
| `summary_of_paper`, `summary_of_recommendation`, `technical_quality`, `clarity_of_presentation`, `potential_impact`, `robotics_focus` | CoRL |
| `extra_scores`, `extra_text` | dicts capturing any venue-specific fields not in the union |
Schema differences in detail:
- **NeurIPS 2021 D&B (track)**: older form, most numeric sub-scores absent; lives in the `neurips` split.
- **NeurIPS 2022–2024**: `soundness/presentation/contribution`; separate `strengths`/`weaknesses`; single general `author_rebuttal`.
- **NeurIPS 2025**: `quality/clarity/significance/originality`; merged `strengths_and_weaknesses`; per-reviewer rebuttals; in-place review revisions tracked via `was_revised` + `final_justification`.
- **ICLR 2020–2026**: NeurIPS-style `soundness/presentation/contribution`; per-reviewer rebuttals.
- **ICML 2025**: similar to NeurIPS-style; some venue-specific fields (e.g. `technical_quality`, `novelty`) live in `extra_scores`.
- **TMLR**: claim-evidence-style structured review; rolling acceptance.
- **EMNLP 2023**: ARR-style review form with `reasons_to_accept`/`reasons_to_reject`/`excitement`/`reproducibility`.
- **CoRL**: robotics-focused review form.
- **COLM**: language-modeling-focused review form.
## Source and collection
- **Source**: [OpenReview](https://openreview.net) — main and any track-level conferences (e.g. NeurIPS Datasets & Benchmarks).
- **Collected**: April 2026 via the [OpenReview Python API](https://github.com/openreview/openreview-py) (a mix of `openreview.api` v2 and legacy v1 for older NeurIPS/ICLR years).
- **Scraping pipeline**: parallel Modal workers (100 containers, single-token reuse to bypass the 3-req/min rate limit). Each forum was fetched with all official reviews, official comments, decision, and author rebuttals; PDFs were downloaded to a Modal volume.
## Markdown / OCR
- **Engine**: [`nvidia/nemotron-ocr-v2`](https://huggingface.co/nvidia/nemotron-ocr-v2)
- **Compute**: 10× NVIDIA L40S GPUs in parallel on Modal (~6 h wall-clock end-to-end)
- **Throughput**: ~7,200 PDFs/hour aggregate; mean **7.3 s/PDF** per worker; **558,785** pages processed
- **Failures**: 1 PDF errored out during OCR; ~76 PDFs were unavailable from OpenReview at scrape time and have `markdown == ""`. Use `markdown_chars > 0` to filter.
### OCR quality (spot-checked across NeurIPS, ICLR, ICML, TMLR, CoRL, COLM)
What works well:
- Body prose, abstracts, section headers, paragraph structure
- In-line citations like `(Author et al., YEAR)` mostly preserved
- Equations rendered linearly (variable names + structure visible)
- Page boundaries marked with `\n\n---\n\n` separator
Recurring artifacts to be aware of (consistent across venues, low impact for most NLP tasks):
- Email addresses and URLs containing repeated chars (e.g. `name@@@cmu.eed`, `https:////aaaaaaa`)
- Occasional word-doubling at line breaks (`decision-decision-making`, `Complex-Valuee Valued`)
- Citation lists missing semicolons (`(Singer 2007 Uhlhaas et al. 2009)`)
- Greek letters and super/sub-scripts often dropped or flattened
- First-page logo/header text occasionally bleeds into title (`git Cooperative …`)
- Figure caption tokens interleave with body text on figure-heavy pages
**Practical impact**: fine for dense retrieval, language modeling, review-grounding, summarization. **Not** suitable for tasks requiring exact equations or canonical citation strings.
## Suggested uses
- Train review-quality classifiers / score predictors
- Study reviewer agreement, rebuttal effectiveness, decision dynamics
- Build retrieval-augmented or grounded scientific-paper assistants
- Meta-research on the peer-review process across venues and years
- Few-shot / RAG benchmarks that require the paper full text + the reviews
## Related work
This dataset was assembled in support of an ICML 2026 Datasets and Benchmarks-track submission introducing **ReviewBench**. Citation will be updated upon publication.
## License
Released under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). Paper full text and review text remain the intellectual property of their respective authors; this dataset redistributes them for non-commercial research consistent with OpenReview's public-access policy. If you are an author and would like content removed, please open an issue on the dataset repository.
## Acknowledgements
- The [OpenReview](https://openreview.net) team for keeping the peer-review record open.
- NVIDIA for releasing [nemotron-ocr-v2](https://huggingface.co/nvidia/nemotron-ocr-v2).
- [Modal](https://modal.com) for the GPU and storage infrastructure used to assemble this corpus.
|