© 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. 22-GEMMASCOPE-RES-131K
    7. 128271
    Prev
    Next
    INDEX
    Explanations

    This neuron detects the visual separators and structured header fields commonly used in appellate‐court briefs—lines of underscores, docket numbers, uppercase metadata tags, and similar all-symbol or all-digit sequences.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    relates to

    np_acts-logits-general · gemini-2.5-flash-lite
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_22/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
     bereits
    -1.30
     nedan
    -1.10
    Ār
    -1.08
    Transkrypt
    -1.01
    letek
    -1.00
     compress
    -1.00
     +(
    -0.97
    🐹
    -0.93
    觑
    -0.93
     ОС
    -0.93
    POSITIVE LOGITS
     امریک
    1.20
     hugely
    1.16
    </em>
    1.14
    —-
    1.11
     colossal
    1.02
     imageNamed
    1.02
     posY
    1.01
     aço
    0.98
     clés
    0.97
     campeón
    0.96
    Activations Density 0.038%

    No Known Activations