© 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. 44593
    Prev
    Next
    INDEX
    Explanations

    proper nouns and specific terms

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

    The neuron strongly activates on uncommon, multi‐syllabic tokens—especially proper names (often with diacritics like “Kōno,” “Tarō,” “Harrell,” “Kieran Timberlake,” “Harlow”) and specialized technical terms (e.g. “hemorrhage”).

    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
    for
    -3.34
    '
    -3.25
    w
    -2.83
    this
    -2.81
    O
    -2.77
    man
    -2.67
    
    -2.66
    no
    -2.53
    All
    -2.44
    No
    -2.38
    POSITIVE LOGITS
    鲻
    2.64
    躥
    2.61
    蹀
    2.55
    嘥
    2.53
    摀
    2.53
    詎
    2.52
     strategia
    2.44
     ventaja
    2.42
    INOIS
    2.41
    轳
    2.39
    Activations Density 0.001%

    No Known Activations