© 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. 19576
    Prev
    Next
    INDEX
    Explanations

    haem-, Lamb, Hägler

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

    Cyrillic characters and scientific notation in technical or medical text.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron is highly sensitive to isolated single‐character “x”/“х” tokens (in both Latin and Cyrillic), i.e. stand-alone occurrences of the letter x.

    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
    ↵↵
    -3.02
    _
    -2.89
    u
    -2.75
    E
    -2.69
    This
    -2.67
    These
    -2.53
    n
    -2.45
    Причины
    -2.42
    b
    -2.41
    t
    -2.33
    POSITIVE LOGITS
    咴
    2.89
    ሷ
    2.70
    鼍
    2.56
    嗩
    2.56
    
    2.53
    Ꮴ
    2.52
     jedną
    2.47
    囯
    2.45
    pecabe
    2.44
    🫅
    2.42
    Activations Density 0.001%

    No Known Activations