© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 118914
    Prev
    Next
    INDEX
    Explanations

    copyright and license

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

    abbreviations or shortened forms of proper names, particularly in technical or formal contexts.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @bfgcn7frf5

    The neuron is sensitive to uncommon subword pieces—rare letter‐sequence fragments (often parts of names, technical terms, or code identifiers) rather than ordinary English words.

    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
    '
    -3.70
     sufrió
    -2.61
     był
    -2.53
     resultó
    -2.08
    父親
    -2.06
    llavero
    -2.05
    桕
    -2.03
     ウィメンズ
    -2.00
    ☫
    -2.00
    是怎么
    -1.95
    POSITIVE LOGITS
     Some
    2.45
     Для
    2.42
     Many
    2.33
    haviours
    2.17
     Некоторые
    2.17
    
    2.14
    vantaged
    2.05
    you
    2.05
    ⦊
    2.05
    fying
    1.98
    Activations Density 0.036%

    No Known Activations