{"paper":{"title":"Survey-to-Behavior: Downstream Alignment of Human Values in LLMs via Survey Questions","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CL","authors_text":"Charles Welch, David Kacz\\'er, Florian Mai, Lucie Flek, Shangrui Nie, Zhixue Zhao","submitted_at":"2025-08-15T11:36:17Z","abstract_excerpt":"Large language models implicitly encode preferences over human values, yet steering them often requires large training data. In this work, we investigate a simple approach: Can we reliably modify a model's value system in downstream behavior by training it to answer value survey questions accordingly? We first construct value profiles of several open-source LLMs by asking them to rate a series of value-related descriptions spanning 20 distinct human values, which we use as a baseline for subsequent experiments. We then investigate whether the value system of a model can be governed by fine-tun"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.11414","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2508.11414/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}