© 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. 22-GEMMASCOPE-RES-131K
    7. 28492
    Prev
    Next
    INDEX
    Explanations

    dialogue starters

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

    This neuron detects the start of lines of dialogue, activating strongly on opening quotation marks and the first word of quoted speech.

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

    the beginnings of quoted dialogue—opening quotation marks, speaker cues, and initial discourse words at the start of a spoken line.

    oai_token-act-pair · gpt-5Triggered by @jyhe0408

    quotation marks or dialogue markers at the beginning of spoken lines in narratives and scripts.

    oai_token-act-pair · claude-4-5-sonnetTriggered 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
     other
    -1.40
     more
    -1.27
     frequently
    -1.20
     there
    -1.20
     less
    -1.16
     otros
    -1.16
     only
    -1.14
     from
    -1.13
     many
    -1.13
     various
    -1.13
    POSITIVE LOGITS
    大丈夫です
    1.33
     seraient
    1.31
    Indeed
    1.23
    Actually
    1.22
     insatiable
    1.21
     intrigu
    1.20
    確かに
    1.20
    ようですが
    1.20
    所谓的
    1.16
     ABSOL
    1.16
    Activations Density 0.005%

    No Known Activations