Steering Black-Box LLMs with Advisor Models

N/A

How to Train Your Advisor: Steering Black-Box LLMs with A DVISOR M ODELS Parth Asawa * 1 Alan Zhu * 1 Abigail O’Neill 1 Matei Zaharia 1 Alexandros G. Dimakis 1 2 Joseph E. Gonzalez 1 Abstract arXiv:2510.02453v3 [cs.LG] 15 May 2026 Frontier language models are deployed as…

related reading

• The Persona Selection Model: Why AI Assistants might Behave like Humansalignment.anthropic.com
• Guardian Angels: LLM Personalization for Productivity and Security · Gwern.netgwern.net
• Composer2.pdfcursor.com
• GenAI Handbookgenai-handbook.github.io
• [2305.18290] Direct Preference Optimization: Your Language Model is Secretly a Reward Modelarxiv.org
• Patterns for Building LLM-based Systems & Productseugeneyan.com
• Model optimization | OpenAI APIplatform.openai.com
• Foundation Models for Oversight | Transluce AItransluce.org
• [2507.19457] GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learningarxiv.org
• Parsed | Custom, interpretable AI systems that continuously learnparsed.com
• Tinkerthinkingmachines.ai
• 2401.10020.pdfarxiv.org