© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 99437
    Prev
    Next
    INDEX
    Explanations

    undesirable conditions and pitfalls

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

    The neuron flags unusually long tokens (e.g. long identifiers or adjectives like “setDefaultCloseOperation,” “customerId,” or “undesirable”).

    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
    糴
    -2.92
    ться
    -2.89
     réservoir
    -2.39
     verlangen
    -2.34
     opérateur
    -2.28
     "
    -2.28
    少し
    -2.25
    だいぶ
    -2.22
     hordes
    -2.22
    老婆
    -2.22
    POSITIVE LOGITS
    .
    3.02
    .“
    2.38
    .”
    2.36
    ,”
    2.36
     was
    2.27
    媜
    2.25
    ,“
    2.17
    án
    2.16
    d
    2.16
     Dezember
    2.14
    Activations Density 0.001%

    No Known Activations