© 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. 25698
    Prev
    Next
    INDEX
    Explanations

    interrogative sentences and uncertainty

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

    The neuron is tuned to high-frequency punctuation and clause‐linking tokens—especially commas and question marks.

    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
    だけど
    -0.96
     nhưng
    -0.93
     lamenta
    -0.92
     tatsächlich
    -0.91
     borracha
    -0.88
    buddha
    -0.87
     certaines
    -0.87
     lyck
    -0.86
     probable
    -0.84
    tierrez
    -0.84
    POSITIVE LOGITS
    ?
    1.52
    Isn
    1.06
    Surely
    1.05
    难道
    1.02
     unless
    1.02
     why
    1.00
    Unless
    0.98
    unless
    0.96
     разве
    0.95
     przecież
    0.95
    Activations Density 0.106%

    No Known Activations