© 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
    2. Gemma-2-27B
    3. 34-GEMMASCOPE-RES-131K
    4. 118300
    Prev
    Next
    INDEX
    Explanations

    design and topology

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

    The neuron strongly activates on formal academic‐paper metadata and technical terms—especially function words (“in,” “for,” “by,” etc.) and math/science keywords (e.g. “variational,” “methods,” “spaces,” “developed”)—as they appear in titles, abstracts, and headings of scholarly articles.

    oai_token-act-pair · o4-miniTriggered 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
    ણ
    -0.76
    ilera
    -0.73
     canceled
    -0.73
    Issued
    -0.72
     Trainers
    -0.69
    erobe
    -0.68
    誰が
    -0.68
    orthand
    -0.67
     Glaser
    -0.67
    issued
    -0.67
    POSITIVE LOGITS
     design
    1.41
    design
    1.20
    DESIGN
    1.15
     optimization
    1.14
     Design
    1.09
    Design
    1.08
     DESIGN
    1.07
     tasarım
    1.04
     topology
    1.03
    デザイン
    1.02
    Activations Density 0.018%

    No Known Activations