© 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. 98664
    Prev
    Next
    INDEX
    Explanations

    asking how something functions or compares

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

    This neuron identifies interrogative question constructions, especially “how”-questions and comparative phrases like “how does …,” “how do …,” “how X stacked up,” or “how X fares.”

    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
     is
    -1.67
     debería
    -1.36
    Oltre
    -1.20
     bahwa
    -1.16
    çois
    -1.14
     Verfasser
    -1.13
     suono
    -1.13
    𝖤
    -1.12
    *
    -1.10
    Before
    -1.09
    POSITIVE LOGITS
     others
    1.69
     against
    1.52
     other
    1.27
    against
    1.27
    umą
    1.26
    iertamente
    1.25
    romeda
    1.24
     rodeado
    1.24
     ...]
    1.23
    middels
    1.19
    Activations Density 0.046%

    No Known Activations