© 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. 34-GEMMASCOPE-RES-131K
    7. 70295
    Prev
    Next
    INDEX
    Explanations

    treat or treatment

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

    The neuron identifies the word "treat" in various contexts, particularly when it appears in programming code (<TreatWarningsAsErrors>), legal documents (treaty), or in phrases like "treat yourself".

    oai_token-act-pair · claude-3-7-sonnet-20250219Triggered by @neilrathi
    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
     aktivi
    -1.23
    ㅏ
    -1.19
    sdf
    -1.09
     rôles
    -1.09
     akut
    -1.09
    "'
    -1.05
     pleads
    -1.02
     angen
    -1.01
     toils
    -1.01
    üsü
    -0.98
    POSITIVE LOGITS
     Treat
    1.77
     treated
    1.72
     treat
    1.70
     Treatment
    1.52
     Treated
    1.51
     behandeln
    1.47
     treats
    1.38
     Treats
    1.32
    treat
    1.27
     behandelt
    1.23
    Activations Density 0.064%

    No Known Activations