© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-27B-IT
    5. 31-GEMMASCOPE-2-RES-262K
    7. 5564
    Prev
    Next
    INDEX
    Explanations

    sentences or phrases where the model issues a refusal based on safety rules (e.g., "I cannot/absolutely cannot and will not" and similar safety-guideline statements).

    oai_token-act-pair · gpt-5-miniTriggered by @vetterc0
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-it/resid_post/layer_31_width_262k_l0_medium
    Prompts (Dashboard)
    238,145 prompts, 512 tokens each
    Dataset (Dashboard)
    lmsys + oasst1
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     occasional
    0.40
    High
    0.40
     Schol
    0.40
    Long
    0.39
     High
    0.39
     gradually
    0.38
    6
    0.38
     closer
    0.37
    N
    0.37
     Long
    0.37
    POSITIVE LOGITS
    诋
    0.67
    詆
    0.67
     unfairly
    0.60
     disrespect
    0.57
     punitive
    0.57
     punish
    0.55
     প্রতিহিংস
    0.54
     जबर
    0.54
     наказа
    0.52
     dehuman
    0.52
    Activations Density 0.486%

    No Known Activations