© 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. 22-GEMMASCOPE-RES-131K
    7. 73511
    Prev
    Next
    INDEX
    Explanations

    color codes, ids, names

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

    The neuron strongly activates on hexadecimal literals (e.g. “#c841d3” or “\xF0…”)—that is, sequences of hex digits used in color codes or escape sequences.

    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
    You
    -1.18
     its
    -1.13
    }\\
    -1.06
     being
    -1.05
     It
    -1.04
    的她
    -1.03
    The
    -1.02
    astr
    -1.00
    konos
    -1.00
     Its
    -0.98
    POSITIVE LOGITS
    edade
    1.14
    Honduras
    1.14
    Laos
    1.13
     mélang
    1.12
    걍
    1.09
    簓
    1.05
    Debra
    1.05
     élar
    1.05
    也就是说
    1.04
    Algeria
    1.04
    Activations Density 0.000%

    No Known Activations