© 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. 103994
    Prev
    Next
    INDEX
    Explanations

    code-like formatting and comments

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

    comment lines or decorative separators in source code (slashes, asterisks, dashes, or whitespace used as visual dividers).

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

    The neuron fires on lines made up of comment or section‐separator punctuation (e.g. “//”, “/* … */”, rows of “—” or “////”), i.e. structural separator or header lines rather than actual code or text.

    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
     with
    -1.79
     -
    -1.50
     including
    -1.32
    biele
    -1.26
     歳
    -1.26
        
    -1.18
     ordinario
    -1.18
     Without
    -1.17
     use
    -1.16
    肫
    -1.16
    POSITIVE LOGITS
    歺
    1.51
     heilig
    1.50
    の中では
    1.48
    pflege
    1.34
     nimmer
    1.27
    resse
    1.26
     JOINT
    1.25
    aning
    1.25
     FORUM
    1.25
    ignment
    1.24
    Activations Density 0.067%

    No Known Activations