© 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. 44356
    Prev
    Next
    INDEX
    Explanations

    high-throughput screens

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

    technical computing terms or coding-related language across different programming contexts.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    This neuron spots tokens in non-Latin scripts (e.g. Chinese characters).

    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
    8
    -3.11
    3
    -2.97
    2
    -2.86
    
    -2.83
    r
    -2.73
    0
    -2.72
    9
    -2.70
    g
    -2.58
    不仅
    -2.44
     这
    -2.33
    POSITIVE LOGITS
     collaboratively
    2.75
     collectively
    2.69
     concurrently
    2.67
     rigorously
    2.61
     become
    2.61
     sincerely
    2.59
     vigorously
    2.50
     あと
    2.47
    狎
    2.45
    ︲
    2.44
    Activations Density 0.001%

    No Known Activations