© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 34-GEMMASCOPE-RES-131K
    7. 700
    Prev
    Next
    INDEX
    Explanations

    of interest or importance

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

    The neuron strongly activates on tokens with very high vocabulary indices (i.e. rare or unusually long/complex subword pieces).

    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
    hungry
    -0.82
    特有
    -0.81
    hastly
    -0.79
    理念
    -0.79
     Stickers
    -0.77
    ">\
    -0.75
     ᾿
    -0.74
     HELEN
    -0.74
     
    -0.73
     Ctr
    -0.73
    POSITIVE LOGITS
     interest
    5.75
    interest
    4.41
     Interest
    4.03
    Interest
    3.88
     concern
    3.70
     interés
    3.47
     importance
    3.27
     interesse
    3.22
     relevance
    3.14
     interests
    3.05
    Activations Density 0.159%

    No Known Activations