© 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. 82503
    Prev
    Next
    INDEX
    Explanations

    polite requests and greetings

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

    This neuron activates on tokens inside quoted speech, effectively detecting lines of dialogue.

    oai_token-act-pair · o4-miniTriggered by @jyhe0408

    dialogue or quoted speech in conversational contexts.

    oai_token-act-pair · claude-4-5-sonnetTriggered by @jyhe0408

    instances of direct speech or dialogue, especially quoted lines and speaker turns marked by punctuation like quotation marks or colons.

    oai_token-act-pair · gpt-5Triggered 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
    メンバー
    -0.96
    thodoxy
    -0.95
    thargy
    -0.93
     reread
    -0.91
    ternut
    -0.91
     goalie
    -0.90
    igrette
    -0.89
     cuando
    -0.89
    ettes
    -0.87
     when
    -0.85
    POSITIVE LOGITS
     gentlemen
    1.30
     Excuse
    1.28
    Excuse
    1.28
     Mr
    1.20
     excuse
    1.16
     gentleman
    1.11
     sir
    1.05
     please
    1.04
     introductions
    1.04
     identification
    1.02
    Activations Density 0.019%

    No Known Activations