VISTA: Steering LLMs with Latent User Values Inferred from a Few Survey Answers
We introduce VISTA (Value-Informed Steering and Targeted Acquisition), which infers a latent profile from users' (full or partial) survey answers, and maps the profile through a hypernetwork to a user-specific steering intervention in a frozen LLM to select answer choices or generate free-text responses.
ProofPaper ↗
Key points
- Personalizing a language model requires knowing what a person values, which stays largely consistent across tasks.
- Given the same evidence, VISTA matches or beats in-context profiles on synthetic preferences (Food2D), the World Values Survey (WVS), and PrimeX, without adding prompt tokens.
- To reduce the cognitive burden of completing full surveys, VISTA also chooses which questions to ask and when to stop: on WVS it stops after 6% of the questionnaire while retaining 94% of the full-history gain, where standard adaptive testing retains at most 42%.
- The dominant steering direction is an interpretable value axis on which VISTA places unseen users accurately.
Sources (1)
- [1]VISTA: Steering LLMs with Latent User Values Inferred from a Few Survey AnswersarXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 16, 04:28 PM
We introduce VISTA (Value-Informed Steering and Targeted Acquisition), which infers a latent profile from users' (full or partial) survey answers, and maps the profile through a hypernetwork to a user-specific steering intervention in a frozen LLM to select answer choices or generate free-text responses.
Personalizing a language model requires knowing what a person values, which stays largely consistent across tasks.
Extractive summary: sentences quoted from the sources.
Before this
- Oct 16, 2026Characterising and Analysing Retrieval-Augmented Generation Techniques
- Oct 15, 2026Decomposing Brain-LLM Alignment with Interpretable Semantic Representations
- Oct 15, 2026Incentivizing Tool-Use Reinforcement Learning for Post-Arrival Emergency Care Prediction
- Oct 15, 2026GlobeBench: Evaluating Text-Based World Model Capabilities for Scalable Environment Simulations
- Oct 15, 2026Librarian Agents: Training LLMs to Curate Long-Context Workspaces that Anticipate Future Queries
- Oct 15, 2026JevAgentHarness: Benchmarking, Deploying and Training Decision Models That Share Control with LLMs