© 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. 113716
    Prev
    Next
    INDEX
    Explanations

    Chat followed by specific words

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

    The neuron fires on the substring “chat” when it appears inside code identifiers (package names, class names, variables, includes, etc.).

    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
    vantaged
    -2.98
    厹
    -2.88
    蠓
    -2.86
    泐
    -2.84
     และ
    -2.70
    咈
    -2.59
    嘸
    -2.55
    ครื่อง
    -2.55
    鸲
    -2.55
    atser
    -2.48
    POSITIVE LOGITS
    y
    3.28
    ar
    3.19
    }
    2.88
     When
    2.80
    `
    2.77
     BOTH
    2.75
    e
    2.73
     unwavering
    2.72
     innumerable
    2.63
    \
    2.63
    Activations Density 0.005%

    No Known Activations