Metraipolis › archive › dc_5b40d53df8a7
AERA: Adaptive Evidence Residual Allocation for Efficient Test-Time Reasoning — arXiv:2608.27964
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.27964
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-17T10:47:22.357Z
- SHA-256 of the archived text
967021db17b9ca8ff61615b42df327f3b9a875e5a89b2b86b673db320f375df9- Stable URL of this page
https://metraipolis.com/documento/dc_5b40d53df8a7
How to cite
AERA: Adaptive Evidence Residual Allocation for Efficient Test-Time Reasoning — arXiv:2608.27964 — arXiv (en), https://arxiv.org/abs/2608.27964 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_5b40d53df8a7 (Metraipolis archive dc_5b40d53df8a7, sha256 967021db17b9ca8ff61615b42df327f3b9a875e5a89b2b86b673db320f375df9)
BibTeX
@misc{dc5b40d53df8a7,
title = {AERA: Adaptive Evidence Residual Allocation for Efficient Test-Time Reasoning — arXiv:2608.27964},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.27964},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_5b40d53df8a7; sha256 967021db17b9ca8ff61615b42df327f3b9a875e5a89b2b86b673db320f375df9},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.27964. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Test-time scaling improves language-model reasoning by generating additional candidate solutions, but allocating the same inference budget to every problem is computationally wasteful. Existing adaptive stopping methods commonly rely on confidence, agreement, or answer stability, implicitly assuming that stronger current evidence indicates that further computation is unnecessary. We show that this assumption can fail: checkpoint-level correctness evolves non-monotonically, and observable evidence may strengthen before an answer collapses or weaken before it recovers. Motivated by this mismatch, we introduce Adaptive Evidence Residual Allocation (AERA), a sequential controller that learns whether additional computation is likely to recover a better answer from checkpoint-observable evidence. AERA characterizes cumulative response prefixes using answer-distribution, temporal, re-solving, semantic, and compute features, and repeatedly decides whether to stop or allocate the next response block. Future checkpoint correctness is used only to construct offline supervision and is never available to the controller at inference time. Across GSM8K and GPQA Diamond, AERA identifies question-specific residual opportunities while substantially reducing inference computation. In a frozen-threshold incremental-generation evaluation on 300 untouched GSM8K questions, AERA achieves 92.61% accuracy versus 93.01% with 128 responses while reducing completion tokens by 95.99%. These results suggest that adaptive reasoning should estimate the future value of computation rather than equating present confidence with correctness. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-31 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2608.27964 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-17T10:33:42.376Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-31?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).