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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-27B
    5. 31-GEMMASCOPE-2-RES-262K
    7. 4447
    Prev
    Next
    INDEX
    Explanations

    refused, abuse, and damage

    np_acts-logits-general · gemini-2.5-flash-lite

    deliberate disruption or refusal

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-27b-pt/resid_post/layer_31_width_262k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     মোটামুটি
    0.30
    涵
    0.30
    சன்
    0.29
    奋
    0.29
     deverá
    0.29
     kindred
    0.29
    阐
    0.29
    顺利
    0.29
    图形
    0.29
     ცხ
    0.29
    POSITIVE LOGITS
     refused
    0.63
     disrespect
    0.63
     disrespectful
    0.61
     maliciously
    0.59
     отказыва
    0.59
     abusive
    0.57
     refus
    0.56
     abuse
    0.56
     refuse
    0.56
     refusing
    0.56
    Activations Density 0.285%

    No Known Activations