# COFFEQA preprint reports gains in small-model question answering

_Published Monday, October 12, 2026 at 6:55 AM EDT · AI, Science · Latest · Tier 2 — Notable_

Researchers report that COFFEQA, a method for training smaller language models with limited examples, improves question answering over training with factual examples alone. In a preprint, they report gains on TriviaQA of 8.77 average exact-match points and 9.98 average F1 points using T5 models and 32 training pairs.

The method pairs factual contexts with minimally modified versions that cause a larger teacher model to give a different answer. It checks that the alternative answer depends on the changed context before using the pairs to train the smaller model. Tests covered five in-domain and three out-of-domain benchmarks using T5 and Qwen2.5 families; the reported advantage is over factual-only distillation at the same few-shot budget.

## Sources

- [cs.LG updates on arXiv.org](https://arxiv.org/abs/2610.12837)

---
Canonical: https://techandbusiness.org/newswire/xCvdm5-r-UrHXsHf0P2_Z-
Published: 2026-10-12T10:55:17.038Z
Story chronology: 2026-10-12T04:00:00.000Z
Retrieved: 2026-10-12T13:18:28.492Z
Publisher: Tech & Business (techandbusiness.org)