© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-3-12B
    5. 31-GEMMASCOPE-2-RES-16K
    7. 3455
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on document section or heading labels (e.g. title-cased words like “Notice,” “Application,” “Suspension,” “Overview,” etc.).

    oai_token-act-pair · o4-mini
    Triggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium
    Prompts (Dashboard)
    392,802 prompts, 256 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    No Configuration Found
    Embeds
    IFrame
    Link
    Negative Logits
    也
    0.83
     begge
    0.80
     बेहद
    0.76
     tüm
    0.72
    但
    0.72
    甚至
    0.71
    狠狠
    0.71
     પણ
    0.67
    或者是
    0.66
     joka
    0.66
    POSITIVE LOGITS
     Revisited
    1.18
     Overview
    1.15
     overview
    1.13
     Considerations
    1.06
     Aspects
    0.98
    และการ
    0.98
     Summary
    0.95
     considerations
    0.94
    Overview
    0.91
     summary
    0.89
    Activations Density 0.185%

    No Known Activations