© 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. 13328
    Prev
    Next
    INDEX
    Explanations

    how and when

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

    The neuron fires on interrogative construction tokens—especially “how” (and, to a lesser extent, “when” and “to”) in question- or instruction-style phrasing.

    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
     sufrió
    -1.91
     falleció
    -1.86
     totalité
    -1.80
     regresó
    -1.80
    
    -1.77
    ゴールデン
    -1.76
    齜
    -1.74
     solidarité
    -1.73
     perdió
    -1.73
     anunció
    -1.72
    POSITIVE LOGITS
    ol
    2.00
    .
    1.90
    s
    1.77
    v
    1.77
     it
    1.64
     how
    1.59
     method
    1.56
    el
    1.54
    ani
    1.53
    ”
    1.52
    Activations Density 0.393%

    No Known Activations