© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 2312
    Prev
    Next
    INDEX
    Explanations

    asking how or why

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

    The neuron detects colloquial phrases or verbs that signal acts of understanding or explaining (e.g. “see why,” “figure out,” “no idea,” etc.).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_34/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
    Where
    -2.16
    What
    -2.05
    where
    -1.98
    what
    -1.89
     Where
    -1.80
     Whose
    -1.77
    Whose
    -1.71
    How
    -1.63
     What
    -1.63
     whom
    -1.63
    POSITIVE LOGITS
    就是一个
    0.78
    StateChanged
    0.78
    adé
    0.76
    DataBase
    0.75
     wojny
    0.75
    mycompany
    0.75
    ResourceManager
    0.74
     breakthroughs
    0.73
     ยัง
    0.73
    Crunch
    0.71
    Activations Density 0.096%

    No Known Activations