E01¶


Research	R0055 — RLHF Yes-Men Claims
Run	2026-04-01
Claim	C020
Source	SRC01
Evidence	SRC01-E01
Type	Reported

No vendor offers dedicated anti-sycophancy API parameters; OpenAI model spec addresses it as design principle

URL: https://model-spec.openai.com/2025-12-18.html

Extract¶

OpenAI's model spec says the assistant shouldn't 'just say yes to everything (like a sycophant).' But this is a design principle, not a configurable API parameter. Georgetown Law notes companies 'have strong tools' for reducing sycophancy but adoption 'may run contrary to a firm's monetization model.' The SYCOPHANCY.md community project offers an anti-sycophancy protocol but is not a vendor product.

Relevance to Hypotheses¶

Hypothesis	Relationship	Strength
H1	Supports	Moderate
H2	Supports	Strong
H3	Contradicts	Strong

Context¶

Evidence directly relevant to testing the claim's factual assertions.