© 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
    2. Gemma-2-27B
    3. 10-GEMMASCOPE-RES-131K
    4. 90866
    Prev
    Next
    INDEX
    Explanations

    that is truly remarkable

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

    The neuron selectively responds to common discourse and linking words—especially high-frequency function words and transitions such as “is,” “for,” and “however.”

    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
     however
    -1.59
     zwar
    -1.42
     però
    -1.39
     respetar
    -1.37
     but
    -1.37
     extremamente
    -1.36
     μπορεί
    -1.35
    both
    -1.35
     sooo
    -1.35
     nhưng
    -1.33
    POSITIVE LOGITS
     from
    1.68
     at
    1.60
     with
    1.47
     for
    1.34
     their
    1.31
    ,”
    1.30
     on
    1.30
     that
    1.29
     other
    1.29
     millon
    1.25
    Activations Density 0.048%

    No Known Activations