© 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. 10-GEMMASCOPE-RES-131K
    7. 84459
    Prev
    Next
    INDEX
    Explanations

    when or first

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

    The neuron detects recommendation or prescription constructions, especially modal verbs like “should” or “must” paired with “be.”

    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
     what
    -1.57
     would
    -1.55
     as
    -1.54
     more
    -1.53
    haviours
    -1.48
     Its
    -1.47
     proven
    -1.41
     many
    -1.37
    neling
    -1.36
    natured
    -1.35
    POSITIVE LOGITS
     Cuando
    1.92
     eerste
    1.67
     your
    1.63
     пър
    1.52
     první
    1.52
     pierwszym
    1.49
     ersten
    1.49
     primeros
    1.47
     primero
    1.45
     когато
    1.44
    Activations Density 0.037%

    No Known Activations