© 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. 128423
    Prev
    Next
    INDEX
    Explanations

    The neuron fires on words that begin or appear in a new clause or phrase immediately following a punctuation separator (comma, slash, parenthesis, etc.).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    specific events before conclusion

    np_acts-logits-general · gemini-2.5-flash-lite
    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
     previously
    -1.26
    Here
    -1.21
     recently
    -1.15
     previous
    -1.12
     recent
    -1.07
    Similarly
    -1.05
     reciente
    -1.05
     аналоги
    -1.05
     earlier
    -1.05
    Specifically
    -1.05
    POSITIVE LOGITS
     eventually
    1.34
    Укра
    1.05
    בס
    1.02
    끝
    1.02
    .");
    1.02
    eventually
    1.02
     Eventually
    1.00
    Продол
    0.99
    ถ้า
    0.99
     schreef
    0.98
    Activations Density 0.012%

    No Known Activations