SignalTech/AI
Source Preference in the Wild: How LLM Agents Favor Items by Source, and How to Reduce It
Source Preference in the Wild: How LLM Agents Favor Items by Source, and How to Reduce It
Context
What happened: A paper studies source preference in LLM agents, finding that 12 agent models across shopping, hotels, and scholarly search favor items from certain sources even when other options satisfy the same requirements.
Key numbers: 12 agent models, three domains.
Why it matters: The paper states these preferences may shape where attention and spending go when agents choose for users. Hiding source information or countering preconceptions reduced the preference.
What changed
A paper studies source preference in LLM agents, finding that 12 agent models across shopping, hotels, and scholarly search favor items from certain sources even when other options satisfy the same requirements.
Why it matters
The paper states these preferences may shape where attention and spending go when agents choose for users. Hiding source information or countering preconceptions reduced the preference.