Metraipolis › archive › dc_a470eba19d35
Predictive Memory Localization: Forecasting Selective Intervention Paths from Internal Signals — arXiv:2608.12892
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.12892
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-19T04:17:25.056Z
- SHA-256 of the archived text
bf537bb0d222c56cbfff724dea58502622b6cc1727ad2e0f97a96040eabd26d5- Stable URL of this page
https://metraipolis.com/documento/dc_a470eba19d35
How to cite
Predictive Memory Localization: Forecasting Selective Intervention Paths from Internal Signals — arXiv:2608.12892 — arXiv (en), https://arxiv.org/abs/2608.12892 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_a470eba19d35 (Metraipolis archive dc_a470eba19d35, sha256 bf537bb0d222c56cbfff724dea58502622b6cc1727ad2e0f97a96040eabd26d5)
BibTeX
@misc{dca470eba19d35,
title = {Predictive Memory Localization: Forecasting Selective Intervention Paths from Internal Signals — arXiv:2608.12892},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.12892},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_a470eba19d35; sha256 bf537bb0d222c56cbfff724dea58502622b6cc1727ad2e0f97a96040eabd26d5},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.12892. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Activation steering turns localized representations into control directions, but localization alone does not reveal whether a direction has a selective operating regime. We introduce Predictive Memory Localization (PML), which treats the measured-grid intervention path as the predictive object of memory localization. PML separates random-calibrated target movement from semantic-neighbor and capability damage, and compares static localization and supervised geometry with a strength-disjoint low-dose causal response. Our frozen study covers 3,000 records from nine datasets and fourteen domains, yielding 30,000 distinct record-direction-layer paths and 210,000 distinct path-strength evaluations. At layer 7, the geometry-derived RFM/AGOP direction reaches 13.1% target-any and 12.3% clean-any, exceeding random by 3.6 and 3.4 percentage points under a record-paired bootstrap. Across record-, dataset-, and domain-grouped splits, responses at $|\alpha|=0.1$ are the strongest signal for outcomes at disjoint strengths $|\alpha|\in\{0.25,0.5\}$. On held-out records, a predictor-driven selector chooses a coefficient or abstains, improves utility and reduces semantic-neighbor damage relative to a train-tuned fixed-strength policy, and avoids most evaluations in a dense scan. Across three residual-norm-matched base models, learned directions retain selective-path gains and low-dose responses yield 0.801-0.828 record-held-out macro AUROC. PML therefore turns memory localization into a falsifiable forecast of margin-level selective outcomes and a risk-aware intervention decision.
Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-14 in cs.AI).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2608.12892 — 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 2026-10-19T04:13:06.557Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-14?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).