PeerQA
We present PeerQA, a real-world, scientific, document-level Question Answering (QA) dataset.
How to run
$ npm i -g lettokenburn-cli$ ltb bench pull peerqa --n 50$ ltb bench run ./peerqa -a claude-code -m anthropic/claude-opus-4-8$ ltb bench push
License is link-only: ltb bench pull fetches rows from the original host for local evaluation only and does not write a solution/ directory. Do not redistribute the generated tasks.
Runs happen on your machine with your own subscription; only the score summary is pushed. CLI docs
Scores
12 sourced entries| Harness | Source | Confidence | Link | |||
|---|---|---|---|---|---|---|
| Mistral-v02-7B-32k | 47.0 | N/A | Independent | medium | Oct 10, 2023 | arxiv.org |
| Command-R-v01-34B-128k | 42.0 | N/A | Independent | medium | N/A | paperswithcode.com |
| GPT-3.5-Turbo-0613-16k | 33.0 | N/A | Independent | medium | May 28, 2020 | arxiv.org |
| Llama-3-IT-8B-8k | 31.1 | N/A | Independent | medium | Jul 31, 2024 | arxiv.org |
| GPT-4o | 30.9 | N/A | Independent | medium | Mar 15, 2023 | arxiv.org |
| Llama-3-IT-8B-32k | 28.8 | N/A | Independent | medium | Jul 31, 2024 | arxiv.org |
| GPT-4o-2024-08-06-128k | 3.5 | N/A | Independent | medium | Mar 15, 2023 | arxiv.org |
| Mistral-v02-7B-32k | 3.4 | N/A | Independent | medium | Oct 10, 2023 | arxiv.org |
| Llama-3-IT-8B-32k | 3.2 | N/A | Independent | medium | Jul 31, 2024 | arxiv.org |
| Llama-3-IT-8B-8k | 3.1 | N/A | Independent | medium | Jul 31, 2024 | arxiv.org |
| Command-R-v01-34B | 3.1 | N/A | Independent | medium | N/A | paperswithcode.com |
| GPT-3.5-Turbo-0613-16k | 3.0 | N/A | Independent | medium | May 28, 2020 | arxiv.org |
Every number links to where it was published. Low-confidence rows are muted and need per-cell verification.
Reproduced by the community
All runsAbout
We present PeerQA, a real-world, scientific, document-level Question Answering (QA) dataset. PeerQA questions have been sourced from peer reviews, which contain questions that reviewers raised while thoroughly examining the scientific article. Answers have been annotated by the original authors of each paper. The dataset contains 579 QA pairs from 208 academic articles, with a majority from ML and NLP, as well as a subset of other scientific communities like Geoscience and Public Health. PeerQA supports three critical tasks for developing practical QA systems: Evidence retrieval, unanswerable question classification, and answer generation. We provide a detailed analysis of the collected dataset and conduct experiments establishing baseline systems for all three tasks. Our experiments and analyses reveal the need for decontextualization in document-level retrieval, where we find that even simple decontextualization approaches consistently improve retrieval performance across architectures. On answer generation, PeerQA serves as a challenging benchmark for long-context modeling, as the papers have an average size of 12k tokens.