© 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-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 30188
    Prev
    Next
    INDEX
    Explanations

    legal and automotive contexts

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

    The neuron consistently fires on formal metadata tokens—section headings, document labels, legal citations, and license‐related terms that appear in headers or metadata blocks.

    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
    ramienta
    -0.77
    pium
    -0.76
    kyrie
    -0.75
    Maur
    -0.73
    glide
    -0.72
    mány
    -0.71
    方には
    -0.71
    yaratan
    -0.71
    ̠
    -0.70
     alér
    -0.70
    POSITIVE LOGITS
     Michigan
    1.33
    Michigan
    1.30
     Detroit
    1.14
    Detroit
    1.10
    Automotive
    0.88
    zieć
    0.87
    :[],
    0.87
     teddy
    0.87
    icial
    0.86
    Ingredienti
    0.86
    Activations Density 0.001%

    No Known Activations