© 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. 112582
    Prev
    Next
    INDEX
    Explanations

    eaten

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

    The neuron strongly activates on the token “eaten,” i.e. it’s a detector for the past-participle “eaten.”

    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
    ’
    -3.53
    ar
    -2.63
    蠖
    -2.53
    俛
    -2.53
    C
    -2.44
     哥
    -2.30
     renowned
    -2.23
     konkret
    -2.20
     ensuing
    -2.14
    爱心
    -2.13
    POSITIVE LOGITS
    
    3.00
    เกง
    2.64
    焄
    2.56
    ꦠ
    2.50
     There
    2.47
     This
    2.39
     använder
    2.39
    не
    2.39
     französische
    2.36
    に
    2.34
    Activations Density 0.002%

    No Known Activations