© 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. 65235
    Prev
    Next
    INDEX
    Explanations

    legal and technical phrasing

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

    The neuron detects high-level discourse markers and transitional phrases (e.g. “methods,” “for example,” “including,” “by contrast,” “more specifically,” “such as,” “generally”) that introduce examples, contrasts, or elaborations in technical or formal 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_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
     šaty
    -1.02
     hlavně
    -1.02
     pomá
    -0.98
     זמן
    -0.97
    逛
    -0.96
     bekommen
    -0.95
    Pantalla
    -0.93
     блюдо
    -0.91
    を使う
    -0.91
    emptions
    -0.91
    POSITIVE LOGITS
     thereto
    1.09
    ėje
    1.06
     materially
    1.06
    )});
    1.03
     pursuant
    1.03
     by
    1.02
     UNDP
    1.00
    inį
    1.00
     hvid
    1.00
     substantially
    0.99
    Activations Density 0.275%

    No Known Activations