© 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. 26838
    Prev
    Next
    INDEX
    Explanations

    life insurancefunctionveinbrushfeaturesestrogenshydroxyapatitespirometry

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

    professional and technical terminology across different fields like medicine, biology, engineering, and law enforcement.

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron consistently fires on domain-specific, formal or technical nouns (e.g. “unit,” “duty,” “pathogen,” “synaptic,” “regulation,” 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_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
    菒
    -1.35
     connecté
    -1.34
    lieving
    -1.18
    쨩
    -1.10
     équipé
    -1.09
     splendid
    -1.09
     absurd
    -1.09
    鮰
    -1.08
     ludicrous
    -1.08
     spectacular
    -1.08
    POSITIVE LOGITS
     their
    1.23
     during
    1.17
     or
    1.13
     after
    1.09
    </h1>
    1.07
     which
    1.03
     through
    1.02
     relatively
    1.02
     around
    1.01
     unrelenting
    1.01
    Activations Density 0.181%

    No Known Activations