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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    2. Gemma-3-12B
    3. 24-GEMMASCOPE-2-RES-16K
    4. 16154
    Prev
    Next
    INDEX
    Explanations

    production of white blood cells, muscle aches, Court of Justice

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

    The neuron fires on words used in formal/legal or official‐statement contexts (e.g. terms, anyone, grounds, victims), i.e. it flags legalistic or administrative language.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    formal, informational phrasing that states rules, conditions, schedules, or service offerings

    oai_token-act-pair · gpt-5Triggered by @jyhe0408

    the word "anyone" or phrases related to legal terms, victims, and formal prohibitions.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_24_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
    '
    0.86
     version
    0.79
     cui
    0.78
    [
    0.75
     finishing
    0.74
     year
    0.73
     winning
    0.73
     style
    0.73
     button
    0.73
    ’
    0.72
    POSITIVE LOGITS
     Baloch
    0.97
     IPTV
    0.93
     MOSFET
    0.93
    يلي
    0.91
     IGBT
    0.90
     ставак
    0.89
    が通販
    0.89
    मियल
    0.87
     پاکستان
    0.86
     стрелецкой
    0.85
    Activations Density 0.000%

    No Known Activations