# EASE-TTT: Evidence-Aligned Selective Test-Time Training for Long-Context Question Answering

Source: [EASE-TTT: Evidence-Aligned Selective Test-Time Training for Long-Context Question Answering](<https://www.sota2.com/research/paper/6a264c9890966ade919bea94>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-06-09T05:00:07.007Z

Authors: Xiaopeng Yuan, Zebin Wang, Suwen Wang, Zongxin Yang, Haohan Wang, Yushun Dong

Published: 2026-06-05T04:49:37.000Z

## Abstract

Long-context question answering (QA) remains challenging for smaller language models even when answer-bearing evidence is already present in the input. Existing within-context retrieval methods localize and expose candidate evidence chunks for the question, but they stop at input-level evidence exposure rather than adapting the query-side attention parameters that control how the model allocates attention over full-context positions. In contrast, lightweight test-time adaptation methods, such as query-only test-time training (qTTT), leave evidence localization unresolved because their generic span-level self-supervised objectives do not identify which context positions support the current answer. In this paper, we propose Evidence-Aligned SElective Test-Time Training (EASE-TTT), a within-context retrieval-augmented test-time training framework that converts selected evidence chunks into a soft attention supervision target over their token positions. Instead of replacing the full context with retrieved chunks, EASE-TTT uses the resulting attention target to guide query-side adaptation, with the adapted model generating the final answer from the original full context. Experiments on six LongBench QA tasks and three small decoder-only language models show that EASE-TTT achieves the strongest macro-average performance among full-context inference, retrieval-only baselines, and qTTT, supporting evidence-aligned test-time adaptation in long-context QA.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Long-context Question Answering on LongBench](<https://www.sota2.com/research/sota/long-context-question-answering-on-longbench.md>): Long-context Question Answering; LongBench; HotPotQA Accuracy: 36.6.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2606.06906>)

[PDF](<https://arxiv.org/pdf/2606.06906.pdf>)