© 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. 10-GEMMASCOPE-RES-131K
    7. 112975
    Prev
    Next
    INDEX
    Explanations

    anyone who

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

    This neuron fires on non‐finite verb phrases (gerunds or infinitives) that signal intended or ongoing actions—e.g. “planning a visit,” “wants to be involved,” “has watched.”

    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
     イラストレーション
    -1.55
    歺
    -1.52
     juu
    -1.44
    嘁
    -1.42
     decidedly
    -1.42
    鉏
    -1.41
    Här
    -1.41
    Dış
    -1.41
     墊
    -1.41
     ſol
    -1.40
    POSITIVE LOGITS
     a
    2.47
     before
    1.98
     their
    1.86
     this
    1.80
     an
    1.74
     should
    1.68
    🥂
    1.64
     which
    1.63
     or
    1.51
     who
    1.48
    Activations Density 0.037%

    No Known Activations