© 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. 15089
    Prev
    Next
    INDEX
    Explanations

    academic citations and programming attributes

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

    The neuron detects specialized scientific or technical terminology, especially chemical‐ or bio‐related compound names.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    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
     provide
    -1.72
     It
    -1.72
     can
    -1.66
     Since
    -1.66
     produces
    -1.65
     M
    -1.65
     at
    -1.64
     provides
    -1.63
     IN
    -1.61
     September
    -1.61
    POSITIVE LOGITS
     intensif
    1.65
     cuit
    1.63
    ほとん
    1.58
     hant
    1.57
     horrid
    1.57
    硨
    1.52
     effici
    1.52
    generally
    1.51
     trev
    1.49
     tred
    1.48
    Activations Density 0.057%

    No Known Activations