© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 43861
    Prev
    Next
    INDEX
    Explanations

    verbs followed by complements

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

    The neuron strongly fires on common grammatical “function” words (e.g. articles, prepositions, conjunctions) rather than content words.

    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
     fordi
    -1.70
     потому
    -1.58
     to
    -1.56
    joyed
    -1.51
    $.
    -1.48
    ḧ
    -1.48
    。
    -1.47
    }$.
    -1.47
     kommune
    -1.46
     mahu
    -1.45
    POSITIVE LOGITS
     are
    2.86
     is
    2.70
     will
    2.06
     should
    1.88
     were
    1.87
     would
    1.83
     must
    1.81
     have
    1.78
     such
    1.66
     could
    1.63
    Activations Density 2.016%

    No Known Activations