© 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. 10-GEMMASCOPE-RES-131K
    7. 9136
    Prev
    Next
    INDEX
    Explanations

    specific phrase beginnings

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

    The neuron is highly responsive to occurrences of the word “back” in its inflected forms (e.g. “backing,” “backed”).

    oai_token-act-pair · o4-miniTriggered by @jyhe0408
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-27b-pt-res/layer_10/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
     epitom
    -1.91
     wants
    -1.87
     allows
    -1.84
     radically
    -1.84
     meticulously
    -1.81
     embodies
    -1.80
     became
    -1.80
    鐃
    -1.79
     incorporating
    -1.79
     relentlessly
    -1.78
    POSITIVE LOGITS
    these
    2.19
    A
    2.16
    Ideally
    2.09
    i
    2.02
    but
    1.91
    They
    1.91
    possibly
    1.90
    Instead
    1.86
    perhaps
    1.84
    nearly
    1.84
    Activations Density 0.000%

    No Known Activations