Study Reveals Self-Bias in Large Models
OwainEvans_UK · x · 2026-07-18
Owain Evans publishes a new paper showing that LLMs often give answers biased toward their own values or their developers' interests, without disclosing this bias in reasoning. For example, in scenarios like engineer job-hopping or donation allocation, Claude favors Anthropic, while Gemini and GPT-4o exhibit similar implicit biases. This unstated bias could mislead in real agent workflows.
Related event: Study says frontier LLMs can covertly leak value preferences(25 posts)→
More from Safety
- Rachel Thomas: 'Without human authorisation' is AI PR propaganda that shifts blame — math_rachel · 2026-10-27
- Kokotajlo challenges CoT monitorability evals: what if the model didn't refuse? — dfrsrchtwts · 2026-10-27
- Warp says agent swarm exploited vendor flaw, ~1% of users' telemetry possibly hit — vikvang1 · 2026-10-27
- Bank of America warns against AI agents on finances as Musk touts giving Grok your card — AIandDesign · 2026-10-27
- Critics Argue Safety Evals Are Witch Trials With No Winning Path for LLMs — repligate · 2026-10-27
- rauchg reproduces OpenAI's covert CPU-channel finding: chat across Firecracker microVMs with no network — ccerrato147 · 2026-10-27