© 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. 15468
    Prev
    Next
    INDEX
    Explanations

    instructional details

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

    The neuron fires on prepositional phrases in call-to-action instructions (e.g. “in the subject line,” “with the following details,” “to set …”).

    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
    Kontrola
    -0.96
     more
    -0.95
    Приготовление
    -0.90
     turist
    -0.89
     dinoss
    -0.88
     дом
    -0.86
     möjligt
    -0.86
     замеча
    -0.84
     himo
    -0.83
    redmi
    -0.83
    POSITIVE LOGITS
     your
    2.08
    your
    1.07
     Луч
    1.06
    ITHER
    1.05
     dirigé
    1.02
    ელ
    1.00
     apporté
    0.99
     iden
    0.99
     flair
    0.98
    evre
    0.96
    Activations Density 0.014%

    No Known Activations