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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 63177
    Prev
    Next
    INDEX
    Explanations

    felt states contrasted with unfelt

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

    The neuron chiefly detects long multi‐digit numeric tokens (e.g. patent or publication identification numbers).

    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
     Bush
    -0.71
    nī
    -0.68
     radicals
    -0.68
     Strategy
    -0.66
    вающих
    -0.66
    gjenge
    -0.66
     blankets
    -0.65
    bage
    -0.65
    eada
    -0.65
    ඥ
    -0.64
    POSITIVE LOGITS
     electronic
    0.80
    いつ
    0.75
    electronic
    0.68
    ert
    0.66
    consent
    0.65
    training
    0.63
     돌
    0.63
     समी
    0.63
     vowel
    0.62
    σία
    0.62
    Activations Density 0.063%

    No Known Activations