© 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. 22-GEMMASCOPE-RES-131K
    7. 92963
    Prev
    Next
    INDEX
    Explanations

    pal followed by specific words

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

    The neuron fires on any occurrence of the substring “pal” (e.g. at the start of words like palindromic, Palmyra, palette, etc.).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    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
    踅
    -2.84
     necesariamente
    -2.59
    Hechos
    -2.50
     This
    -2.42
     apparent
    -2.39
    脔
    -2.39
    見ると
    -2.38
     personalmente
    -2.33
     regularmente
    -2.31
    錞
    -2.27
    POSITIVE LOGITS
    linawan
    2.69
    ază
    2.64
    鳎
    2.63
     είναι
    2.41
    ሖ
    2.39
    ТА
    2.38
    ágenes
    2.36
            
    2.36
     аккура
    2.36
    解决
    2.34
    Activations Density 0.016%

    No Known Activations