© 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. 77280
    Prev
    Next
    INDEX
    Explanations

    wide followed by other words

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

    The neuron strongly activates on the word “wide” (and its capitalized form “Wide”) when used as a modifier in technical or formal term phrases.

    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
     go
    -0.88
     smaller
    -0.88
     to
    -0.84
     websites
    -0.83
     finding
    -0.82
    ないように
    -0.81
     guided
    -0.81
     more
    -0.81
    orgung
    -0.81
     finer
    -0.80
    POSITIVE LOGITS
    anging
    1.31
    Wide
    1.30
     Wide
    1.21
     wide
    1.21
    awake
    1.17
     WIDE
    1.16
    band
    1.10
     répand
    1.09
     språk
    1.09
     awake
    1.06
    Activations Density 0.041%

    No Known Activations