© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-27B
    5. 10-GEMMASCOPE-RES-131K
    7. 43936
    Prev
    Next
    INDEX
    Explanations

    designated or likely

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

    The neuron is detecting the relative‐clause introducer “which” (and similar function words that start subordinate clauses).

    oai_token-act-pair · o4-miniTriggered 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
    Not in Any Lists

    No Comments

    Negative Logits
     of
    -1.83
     momentan
    -1.51
     verra
    -1.51
     are
    -1.48
     pengar
    -1.47
     relais
    -1.47
     pittores
    -1.45
     toilette
    -1.45
    return
    -1.44
     garan
    -1.41
    POSITIVE LOGITS
     that
    2.75
     causing
    1.70
     including
    1.67
     when
    1.65
     specially
    1.59
     unique
    1.59
     designed
    1.58
     tailored
    1.58
     new
    1.55
     complicated
    1.55
    Activations Density 0.110%

    No Known Activations