E01¶


Research	R0056 — RLHF Yes-Men Claims v2
Run	2026-04-01
Claim	C004
Source	SRC01
Evidence	SRC01-E01
Type	Reported

Primary evidence for C004

URL: See source scorecard

Extract¶

The specific 84-85 percent reduction figure could not be located in any referenced papers. The 84 percent figure in Anthropic paper refers to model knowledge of misconceptions, not sycophancy reduction from preference pairs.

Relevance to Hypotheses¶

Hypothesis	Relationship	Strength
H1	Supports	See assessment
H2	Supports	See assessment
H3	Contradicts	See assessment

Context¶

See assessment.md for full context.