© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 58135
    Prev
    Next
    INDEX
    Explanations

    attributing speech to a speaker

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

    The neuron fires on journalist attribution phrases—particularly the reporting verb “said” (and similar speaker-tag constructions) indicating who’s speaking.

    oai_token-act-pair · o4-mini
    Triggered 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
    Negative Logits
     for
    -1.30
     reportedly
    -1.20
     geração
    -1.16
     sınır
    -1.13
     například
    -1.12
     always
    -1.12
    Matricula
    -1.12
     to
    -1.09
     Schriftsteller
    -1.09
     gestão
    -1.08
    POSITIVE LOGITS
    DCT
    1.30
     its
    1.24
     fi
    1.23
     coreano
    1.22
     два
    1.22
     decorativo
    1.20
    Italie
    1.20
    ۉ
    1.19
     hadapan
    1.17
     are
    1.16
    Activations Density 0.018%

    No Known Activations