© 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. 10-GEMMASCOPE-RES-131K
    7. 21500
    Prev
    Next
    INDEX
    Explanations

    hexadecimal numbers

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

    The neuron is detecting hexadecimal byte tokens (i.e. 1–2-digit hex values) in a hex dump or similar data listing.

    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
     came
    -2.78
    .’
    -2.72
     believes
    -2.58
     декабря
    -2.58
     went
    -2.39
     took
    -2.38
     told
    -2.34
     keeps
    -2.33
     began
    -2.33
    le
    -2.33
    POSITIVE LOGITS
    often
    2.88
    4
    2.86
    8
    2.73
    9
    2.70
    ↵
    2.67
    usually
    2.63
    mainly
    2.61
    すごく
    2.59
    AN
    2.55
     dritte
    2.55
    Activations Density 0.016%

    No Known Activations