Metraipolis › archive › dc_f843dfe92c2d
Conflict and Congruency Effects in Large Language Models: In-Weight and In-Context Competition in a Verbal Conflict Task — arXiv:2608.11510
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.11510
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-19T05:15:54.339Z
- SHA-256 of the archived text
1985643a905891300946fe21d62fe0a4493d6012d6d009eab5c73d434a3d0c38- Stable URL of this page
https://metraipolis.com/documento/dc_f843dfe92c2d
How to cite
Conflict and Congruency Effects in Large Language Models: In-Weight and In-Context Competition in a Verbal Conflict Task — arXiv:2608.11510 — arXiv (en), https://arxiv.org/abs/2608.11510 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f843dfe92c2d (Metraipolis archive dc_f843dfe92c2d, sha256 1985643a905891300946fe21d62fe0a4493d6012d6d009eab5c73d434a3d0c38)
BibTeX
@misc{dcf843dfe92c2d,
title = {Conflict and Congruency Effects in Large Language Models: In-Weight and In-Context Competition in a Verbal Conflict Task — arXiv:2608.11510},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.11510},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f843dfe92c2d; sha256 1985643a905891300946fe21d62fe0a4493d6012d6d009eab5c73d434a3d0c38},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.11510. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Congruency effects, observed in conflict tasks such as Stroop and flanker tasks, have been investigated for nearly a century in psychology and neuroscience, but their mechanistic basis is not fully understood. We introduce a verbal-only LLM conflict task in which a prompt stem elicits a default same-color completion and an explicit rule either agrees with (congruent condition) or conflicts with (incongruent condition) the completion. Gemma-2-2B and six Pythia models ranging from 410M to 12B parameters showed strong default same-color tendencies, and six of seven models showed strong congruency effects. Using causal attribution analysis, attention analysis, and attention ablations, we identified distinct processing pathways in these LLMs: a pathway involving short-range attention to a superficial color cue that is preferentially activated in the congruent condition, and a pathway involving long-range attention to the rule prefix that is preferentially activated in the incongruent condition. Fine-tuning that strengthened the default same-color tendency had divergent effects on task conditions, reducing incongruent performance while increasing congruent performance. In contrast, increasing rule set size selectively impaired incongruent performance. These converging findings support an account in which congruency effects in this task arise from competition between an in-weight default mapping and an in-context rule-based mapping. More broadly, our findings illustrate how LLMs can serve as model systems for mechanistic analysis of competition between default and rule-governed response tendencies within a single learned network. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-13 in cs.AI). Categorie arXiv: Neurons and Cognition (q-bio.NC) ; Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2608.11510 — 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 2026-10-19T05:08:46.693Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-13?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).