© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 48284
    Prev
    Next
    INDEX
    Explanations

    content descriptions and details

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

    This neuron fires on document structural labels or headings (e.g. “Event,” “Review,” “Video,” “Photo,” “Location,” “Summary,” etc.), i.e. metadata field or section-title tokens.

    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
     toiture
    -1.05
    Ingrediente
    -1.04
    Neden
    -0.97
     Anschließend
    -0.97
    独特的
    -0.97
    ického
    -0.96
     Norsk
    -0.96
     gebeten
    -0.94
    
    -0.93
    はじめまして
    -0.93
    POSITIVE LOGITS
     details
    1.75
     description
    1.74
     Description
    1.41
     type
    1.41
     name
    1.39
     length
    1.31
     also
    1.28
     Details
    1.25
     is
    1.24
     not
    1.23
    Activations Density 0.260%

    No Known Activations