Metraipolis › archive › dc_fb5a7ea7c744

Tracing How Language Models Integrate Temporal Anchors into Factual Recall — arXiv:2611.01945

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.01945
License (as archived)
CC0-1.0
Archived
2026-11-03T11:10:39.542Z
SHA-256 of the archived text
206ccc9c05990199f330c8c5142c60e45e7c8f7eee515768f7ae98cb9208e982
Stable URL of this page
https://metraipolis.com/documento/dc_fb5a7ea7c744

How to cite

Tracing How Language Models Integrate Temporal Anchors into Factual Recall — arXiv:2611.01945 — arXiv (en), https://arxiv.org/abs/2611.01945 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_fb5a7ea7c744 (Metraipolis archive dc_fb5a7ea7c744, sha256 206ccc9c05990199f330c8c5142c60e45e7c8f7eee515768f7ae98cb9208e982)
BibTeX
@misc{dcfb5a7ea7c744,
  title = {Tracing How Language Models Integrate Temporal Anchors into Factual Recall — arXiv:2611.01945},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.01945},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_fb5a7ea7c744; sha256 206ccc9c05990199f330c8c5142c60e45e7c8f7eee515768f7ae98cb9208e982},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.01945. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Factual recall in language models is typically framed as retrieving a factual association: a subject--relation pair maps to a single object. However, many facts are context-dependent: the same pair maps to different objects depending on a condition such as time. We introduce anchored factual recall, in which the model receives an explicit conditioning expression, an anchor, and must select the object valid under it. This raises questions that single-object recall does not: where the anchor enters the computation, what it changes, and how that change selects the answer. Focusing on temporal anchors, we answer them with causal interventions on four open-weight models on English prompts and identify two routes. In a subject-mediated route, the anchor resolves the final subject-token representation to the queried year, and the relation redistributes the readout of this state across attention heads. In a direct route] the final position attends to the anchor, and upper-layer MLPs turn this signal into a preference for the time-appropriate object. The two routes are parallel channels: their contributions add up, most queries use both, and their balance depends on the query, e.g., on the relation, consistently across models. This reframes mechanistic accounts of factual recall from retrieving a single stored object to selecting among context-dependent alternatives.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2611.01945 — 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 2026-11-03T10:59:52.318Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).