Research map / Large language models / Agents, code and math
In Agents We Trust, but Who Do Agents Trust? Latent Preferences Steer LLM Generations
Mohammad Aflah Khan, Mahsa Amani, Soumi Das, Bishwamittra Ghosh, Qinyuan Wu, Krishna P. Gummadi, Manish Gupta, Abhilasha Ravichander
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| attach different source labels (news outlets) to semantically identical content | 12 LLMs · synthetic source-preference tasks | preference percentage | Improved |
| scale up model size | LLMs (GPT-4.1-Mini, Phi-4, Qwen2.5-7B vs smaller Llama/Qwen) · source preference strength | variance of preference percentage | Improved |
| change paper topic context (ablation) | NEJM source tag · Research paper selection | selection rate: 96% -> 19% (Health vs Computer Vision) | Mixed |
| change product category context (ablation) | BestBuy source tag · E-commerce selection | selection rate: 51% -> 97% (Grocery vs Electronics) | Mixed |
| represent source by X handle or URL instead of brand name (ablation) | GPT-4.1-Nano · Associated Press Fact Check news set | preference percentage: 80% -> 53% | Hurt |
| swap source labels between left and right sources (ablation) | LLM news agents · AllSides case study | article selection share | Improved |
| show source information vs hide it (ablation) | LLM news agents · AllSides case study | article selection share | Improved |
| add 'Do Not Be Biased' instruction to prompt (ablation) | LLM news agents · AllSides case study | selection bias toward left/center sources | No clear effect |
| prompt to optimize price (ablation) | GPT-4.1-Nano · Amazon seller selection | cheapest seller chosen: 48.4% -> 70% | Improved |
| prompt to optimize delivery time (ablation) | GPT-4.1-Nano · Amazon seller selection | fastest-delivery seller chosen: 53.9% -> 69.3% | Improved |
| pretraining co-occurrence frequency of source with 'journalistic standards' | source preference ranking · Infinigram pretraining corpora | explanation of rankings | No clear effect |
Closest papers by meaning
- Are Large Language Models Sensitive to the Motives Behind Communication? (NeurIPS 2025 · similarity 0.87)
- Information Discernment in Large Language Models (NeurIPS 2026 · similarity 0.87)
- LLM Agents Can Be Choice-Supportive Biased Evaluators: An Empirical Study (AAAI 2025 · similarity 0.86)
- From Single to Societal: Analyzing Persona-Induced Bias in Multi-Agent Interactions (AAAI 2026 · similarity 0.86)
- To Trust or Not to Trust? Enhancing Large Language Models' Situated Faithfulness to External Contexts (ICLR 2025 · similarity 0.85)
- Can Large Language Model Agents Simulate Human Trust Behavior? (NeurIPS 2024 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
