© 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. 34-GEMMASCOPE-RES-131K
    7. 49780
    Prev
    Next
    INDEX
    Explanations

    what are you asking about

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

    The neuron fires strongly on directive or instructional language—words laying out tasks, steps, or assignments (e.g. “pay,” “attention,” “consider,” “bring,” “provide,” 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_34/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
     herbes
    -1.75
     gena
    -1.74
     gela
    -1.72
     simpel
    -1.63
     prins
    -1.61
     tirs
    -1.60
     vernac
    -1.59
     boya
    -1.57
     fatale
    -1.57
     tille
    -1.55
    POSITIVE LOGITS
     or
    2.53
     your
    2.41
     you
    2.34
     each
    1.68
     might
    1.66
     this
    1.60
     would
    1.60
     seemed
    1.56
     similar
    1.56
     different
    1.55
    Activations Density 0.054%

    No Known Activations