© 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. 25482
    Prev
    Next
    INDEX
    Explanations

    questions starting with Q:

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

    The neuron detects cooking recipe steps—especially imperative cooking action verbs (e.g. “Slice,” “Combine,” “Mix,” “Rub”).

    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
     we
    -1.30
     this
    -1.29
     these
    -1.22
     этих
    -1.05
     those
    -1.03
    _
    -0.98
     it
    -0.97
     Weiterbildung
    -0.97
     impresión
    -0.97
     their
    -0.96
    POSITIVE LOGITS
    ？？
    1.28
    ？？？
    1.28
    ！！！！
    1.16
    🫶
    1.13
     polaire
    1.08
    ♡♡
    1.07
     vocale
    1.02
    brada
    1.02
    🤤
    1.02
     folle
    1.01
    Activations Density 0.005%

    No Known Activations