© 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. 73384
    Prev
    Next
    INDEX
    Explanations

    as to why/who/how/where/whether

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

    This neuron strongly activates on the function word “to,” especially when marking infinitives or in phrases like “as to” or “how to.”

    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
    ᥫ
    -1.99
    鯪
    -1.86
    ꗃ
    -1.73
    𓄹
    -1.70
    麪
    -1.65
    晅
    -1.63
    ,
    -1.59
     parachoque
    -1.59
    𓋼
    -1.59
    敍
    -1.58
    POSITIVE LOGITS
     why
    1.93
     当
    1.84
     REALLY
    1.70
     `
    1.58
     whether
    1.50
     WHEN
    1.44
    really
    1.43
     —
    1.41
     quando
    1.38
    os
    1.37
    Activations Density 0.022%

    No Known Activations