---
license: cc-by-4.0
language:
- en
pretty_name: Nexa Science Multitask Balanced
task_categories:
- text-classification
- question-answering
size_categories:
- 10K<n<100K
---

# Nexa Science Multitask Balanced

This dataset is a curated, instruction-formatted scientific multitask mixture for:
- claim verification (`<TASK:VERIFY>`)
- abstract-grounded biomedical QA (`<TASK:QA>`)
- retrieval relevance re-ranking (`<TASK:RERANK>`)

## Format
Each row is JSONL with:
`{task, instruction, input, output, meta}`

## Splits Included
- `train_balanced_short.jsonl`
- `val_balanced_short.jsonl`
- `stats_balanced_short.json`

## Notes
- QA in this balanced release is label-balanced (yes/no/maybe).
- Re-ranking is provided as coarse relevance classes (0-3).
- This release is intended for multitask adapter training and evaluation.

HF repo: https://huggingface.co/datasets/Allanatrix/nexa-science-multitask-balanced