© 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-2-27B
    5. 22-GEMMASCOPE-RES-131K
    7. 128977
    Prev
    Next
    INDEX
    Explanations

    The neuron flags very long, information‐dense text spans (i.e. lengthy technical or explanatory sentences).

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

    easy access

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     StringBuilder
    -0.91
     wherein
    -0.89
    pherical
    -0.83
    閂
    -0.83
     }(
    -0.83
    startIndex
    -0.81
    Alpha
    -0.81
    חד
    -0.81
     можуть
    -0.80
    ograd
    -0.80
    POSITIVE LOGITS
    ESTY
    1.00
     effortless
    1.00
     cumplido
    0.93
     then
    0.93
    Caution
    0.93
     or
    0.93
    CAUTION
    0.92
     tegens
    0.91
    toHaveBeen
    0.87
     Зачем
    0.87
    Activations Density 0.042%

    No Known Activations