© 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. 10-GEMMASCOPE-RES-131K
    7. 39174
    Prev
    Next
    INDEX
    Explanations

    names and quotes

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

    This neuron strongly activates on the first few tokens of a new article or paragraph—i.e. the opening words (often in headlines or lead sentences) and the very first punctuation or quote that marks the start of a news segment.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     his
    -1.48
     such
    -1.29
     różnych
    -1.28
     brinda
    -1.23
     redor
    -1.21
     tohoto
    -1.21
    ...........
    -1.20
     selepas
    -1.20
     schöner
    -1.20
     různých
    -1.16
    POSITIVE LOGITS
    迥
    1.30
     bieber
    1.19
     jose
    1.19
     himo
    1.17
     rita
    1.16
     cherokee
    1.16
     sebastian
    1.13
     στι
    1.13
     Pantal
    1.13
    yoruz
    1.13
    Activations Density 0.058%

    No Known Activations