© 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. 22-GEMMASCOPE-RES-131K
    7. 70944
    Prev
    Next
    INDEX
    Explanations

    code comments and specific entities

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

    The neuron strongly activates on JavaScript/Redux code tokens—especially keywords and identifiers inside an async try/catch (e.g. “try”, “catch”, “dispatch”, braces, colons) rather than on ordinary prose.

    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
     these
    -0.97
    sure
    -0.94
     Estos
    -0.94
    size
    -0.93
     ?";
    -0.88
     condem
    -0.86
    dress
    -0.86
     растение
    -0.85
     vilka
    -0.85
    歉
    -0.84
    POSITIVE LOGITS
    我們
    1.06
     zamiast
    1.00
     chiese
    0.94
    История
    0.93
    เรา
    0.93
    Опис
    0.92
    estroying
    0.91
     hefyd
    0.90
    versy
    0.89
    我們的
    0.88
    Activations Density 0.001%

    No Known Activations