© 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. 81068
    Prev
    Next
    INDEX
    Explanations

    file names and labels

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

    This neuron detects compound identifier tokens (especially hyphen-separated icon or image names) such as “glass-martini,” “meet-up,” “video-play,” “suit-case,” and “star-o.”

    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
     beweg
    -0.81
     February
    -0.79
     molta
    -0.75
     LOTS
    -0.74
     dovol
    -0.74
     glutton
    -0.73
     Pretty
    -0.72
     October
    -0.72
     empiezan
    -0.72
     당신
    -0.71
    POSITIVE LOGITS
    alt
    1.03
     ImageIcon
    0.89
    Alt
    0.87
     offrent
    0.85
    OOOO
    0.82
    ĥ
    0.81
    加以
    0.81
     remplir
    0.80
     DEFENSE
    0.80
    *}$
    0.79
    Activations Density 0.007%

    No Known Activations