© 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. 16346
    Prev
    Next
    INDEX
    Explanations

    basis for determining

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

    The neuron activates on single words that serve as the core relational or “pivot” terms in explanatory statements—high-importance connectives and verbs like “key,” “to,” “for,” “accounts,” “determines,” 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
    pography
    -1.23
    īdz
    -1.04
    ifying
    -1.02
     spør
    -1.01
    ícias
    -1.01
     NDL
    -1.00
     allerede
    -1.00
     ALREADY
    -1.00
    querque
    -0.98
    hæng
    -0.98
    POSITIVE LOGITS
     most
    1.68
     why
    1.45
     successful
    1.35
     many
    1.29
     all
    1.22
     our
    1.20
     success
    1.17
     every
    1.15
     entire
    1.10
     so
    1.08
    Activations Density 0.089%

    No Known Activations