© 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. 34-GEMMASCOPE-RES-131K
    4. 34943
    Prev
    Next
    INDEX
    Explanations

    how an action is done

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

    The neuron fires on manner-or-intensity modifiers (chiefly adverbs and adjectives describing how something is done, e.g. “hard,” “deeper,” “slowly,” “quickly”).

    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
     that
    -1.09
     a
    -0.96
     entirely
    -0.94
     heavily
    -0.92
     сильно
    -0.91
     outright
    -0.90
     more
    -0.88
     acted
    -0.87
     dergleichen
    -0.84
    にいる
    -0.83
    POSITIVE LOGITS
    aneously
    1.41
    ingly
    1.36
    quely
    1.31
    istically
    1.25
    tively
    1.18
    ishly
    1.16
    matically
    1.12
    ulously
    1.09
    ably
    1.06
    ly
    1.03
    Activations Density 0.087%

    No Known Activations