© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 111800
    Prev
    Next
    INDEX
    Explanations

    purpose or finding

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

    This neuron fires on verbs and phrases that introduce explanations or key points (e.g. “发现”, “means”, “展示”, “建议”), i.e. discourse markers signaling summaries or explanations.

    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
     or
    -2.33
     ("
    -2.28
    i
    -2.25
    -"
    -2.23
     "
    -2.19
     on
    -2.16
    鋮
    -2.06
     a
    -1.92
    かと思います
    -1.92
    ="
    -1.91
    POSITIVE LOGITS
     they
    2.08
    ⩴
    2.05
    盌
    1.97
     there
    1.91
    腘
    1.88
    蛱
    1.76
    ῞
    1.74
    whose
    1.72
    氿
    1.72
     hunne
    1.70
    Activations Density 0.005%

    No Known Activations