© 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. 9357
    Prev
    Next
    INDEX
    Explanations

    scientific methods and applications

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

    The neuron identifies phrases where "and" appears in academic article titles, particularly in the structure "topic: [subject] and [subject]".

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi

    The neuron strongly detects the coordinating conjunction “and” when it links two key technical or thematic terms (as seen in paper titles or section headings).

    oai_token-act-pair · o4-miniTriggered 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
     dini
    -1.16
    猫咪
    -1.11
     smör
    -1.08
     etc
    -1.08
     vokal
    -1.05
    酲
    -1.05
     Sì
    -1.04
     kemer
    -1.03
     negatif
    -1.03
    lorn
    -1.03
    POSITIVE LOGITS
     consequences
    1.22
     how
    1.17
     strategies
    1.15
     practical
    1.11
     proceso
    1.10
    门外
    1.09
     worldwide
    1.09
     invitaciones
    1.06
     strategy
    1.05
    INGREDIENTES
    1.05
    Activations Density 0.044%

    No Known Activations