© 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-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 13214
    Prev
    Next
    INDEX
    Explanations

    illegal activities and healthy relationships

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

    The neuron fires on words that denote wrongdoing or illegality (e.g. “recklessly,” “unlawful,” “illegal,” “improper”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_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
     offend
    0.66
     wrongdoing
    0.65
     保護
    0.60
    thebetter
    0.59
     bigotry
    0.59
    錯
    0.59
     sinned
    0.58
    错
    0.58
     защи
    0.57
     सबूत
    0.57
    POSITIVE LOGITS
     exuber
    0.76
    practices
    0.76
     practices
    0.75
     behaviors
    0.71
     Practices
    0.69
     behavior
    0.68
     excessive
    0.68
     yere
    0.65
    findings
    0.64
     Behavior
    0.63
    Activations Density 0.169%

    No Known Activations