© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 60285
    Prev
    Next
    INDEX
    Explanations

    you can use

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

    The neuron fires on the opening words of standalone instructional or policy‐style sentences (e.g. “If,” “By,” “Note,” “You,” “To,” etc.), effectively detecting the start of a new directive or setting explanation.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/width_131k
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    茈
    -1.98
    seems
    -1.88
    extremely
    -1.88
     semblent
    -1.85
    essentially
    -1.72
    somewhat
    -1.70
    customized
    -1.69
    Extremely
    -1.65
    couldn
    -1.65
    
    -1.64
    POSITIVE LOGITS
     with
    2.42
     in
    2.23
     without
    2.17
     on
    2.11
     will
    1.95
     you
    1.94
     or
    1.84
     while
    1.79
     during
    1.73
     if
    1.70
    Activations Density 0.192%

    No Known Activations