© 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
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 28916
    Prev
    Next
    INDEX
    Explanations

    technical description

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

    The neuron fires on non-natural-language tokens typical of code or technical snippets—e.g. punctuation, symbols, numbers, all-caps identifiers, version strings, etc.

    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
     accol
    -1.42
     dollari
    -1.34
    <!--
    -1.34
    ⁉
    -1.34
     ?";
    -1.30
     soos
    -1.30
     distru
    -1.30
     abband
    -1.30
     aperto
    -1.29
    reportWebVitals
    -1.28
    POSITIVE LOGITS
     these
    1.89
    これらの
    1.67
     then
    1.48
     unterscheiden
    1.41
     zusätz
    1.41
    רושלים
    1.40
     entscheid
    1.38
     These
    1.38
     hingegen
    1.38
     wählen
    1.38
    Activations Density 0.029%

    No Known Activations