© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 104290
    Prev
    Next
    INDEX
    Explanations

    \u , hex codes

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

    escape sequences and special characters in code, particularly backslash notation for unicode or hexadecimal character codes.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @bfgcn7frf5

    The neuron is detecting hexadecimal digits (0–9 and A–F) in escape‐sequence contexts.

    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
     about
    -2.19
     which
    -1.80
     then
    -1.63
     whose
    -1.38
     a
    -1.34
     soñ
    -1.32
    cerita
    -1.32
    だったので
    -1.31
     których
    -1.30
    ↵
    -1.29
    POSITIVE LOGITS
    You
    1.95
    Whether
    1.64
    siasme
    1.58
    If
    1.53
    There
    1.52
    Because
    1.51
    Remember
    1.51
    According
    1.48
    While
    1.47
    derdag
    1.46
    Activations Density 0.005%

    No Known Activations