© 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
    2. Gemma-2-27B
    3. 34-GEMMASCOPE-RES-131K
    4. 2928
    Prev
    Next
    INDEX
    Explanations

    rationally related classification

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

    The neuron is strongly triggered by the structural tokens of court‐case citations and headings—numbers, parentheses, reporter abbreviations, court names, and docket references that appear in legal‐case headers.

    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
    ignal
    -0.90
     Pinus
    -0.86
     %#
    -0.77
    limate
    -0.76
    otum
    -0.76
     Bianco
    -0.75
    ohyd
    -0.74
    🍅
    -0.73
    صد
    -0.73
    ppus
    -0.73
    POSITIVE LOGITS
     Classi
    1.12
     Rational
    1.08
    Classi
    1.02
     classification
    1.01
     rational
    0.99
     classifications
    0.98
     Classification
    0.87
    classi
    0.87
     classific
    0.87
     challenger
    0.85
    Activations Density 0.042%

    No Known Activations