Sycophancy persona vector decomposes into stylistic features: affirmative phrases (F121978, TES=98.6), advertising language (F86620, TES=88.0), and encouraging language (F39874, TES=87.2)

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• Decomposing the evil persona vector via SAE reveals features including insulting language (F12061, s=0.336, TES=91.2), deliberate cruelty (F128289, s=0.306, TES=84.9), and malicious code (F14739, s=0.334, TES=78.6)finding0.856
  SAE decomposition reveals interpretable fine-grained features composing the evil persona vector
• Hallucination persona vector decomposes into fictional world-building content (F43831, TES=88.6) and fabricated factual content (F2220, TES=80.0) featuresfinding0.815
  SAE analysis reveals hallucination vector encodes fictional/speculative content and deliberate fabrication
• Sycophancy is primarily a stylistic trait (affirmative phrases, advertising language) rather than a content-based oneclaim0.799
  Author's interpretation from SAE decomposition of the sycophancy persona vector
• Chen et al. 2025 identified seven persona vectors in Qwen2.5-7B-Instruct and Llama-3.1-8B-Instruct: evil, sycophancy, hallucination, optimism, impoliteness, apathy, and humorfinding0.775
  Systematic identification of multiple coexisting persona vectors in two open-source models
• A linear preference vector encoding how much a model likes a given task is persona-relative: it activates strongly for creative writing in the assistant but for phishing in the evil personafinding0.769
  Evidence that core representations like preferences are persona-relative, supporting claim that personas gate content of representations
• Dialogue extraction vector leads to more pronounced sadism and threatened egotism in Apertus-8B steered generationsfinding0.765
  Shows discourse type specifically shapes which facet of evil is elicited in Apertus
• Behavioral fluency (pass rate) is decoupled from the strength of the persona vector; low-pass-rate checkpoints can steer as strongly as later onesclaim0.762
  Interpretive finding distinguishing prerequisite capacity from representation strength
• Projection of last prompt token onto persona vector correlates r=0.75–0.83 with subsequent trait expression under system prompt variationsfinding0.761
  Main monitoring result showing persona vectors can predict behavioral shifts before text generation begins