© 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. 68801
    Prev
    Next
    INDEX
    Explanations

    dates and versions

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

    The neuron fires on metadata and structural markers in a document—things like version headers, abstract/bibliography tags, “Forthcoming,” “comes,” and other tokens that occur in paper preambles or section‐heading lines.

    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
    mnop
    -0.91
    FULLY
    -0.89
    ッピー
    -0.83
    waarde
    -0.83
    în
    -0.83
    ASTM
    -0.83
     Retrospective
    -0.83
     Cardano
    -0.82
     Poincar
    -0.82
     куртка
    -0.82
    POSITIVE LOGITS
     version
    1.22
    Übernach
    1.05
     this
    1.01
    Gründe
    0.96
    UNTAIN
    0.96
     rédig
    0.90
    ormais
    0.88
    版本
    0.88
    などなど
    0.87
     این
    0.86
    Activations Density 0.000%

    No Known Activations