© 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 EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Gemma-2-2B
    5. 6-GEMMASCOPE-TRANSCODER-16K
    7. 14515
    Prev
    Next
    INDEX
    Explanations

    This neuron appears to be firing for a wide variety of different words. Given the high activations for `hard`, `works`, `part-time` and `job`; and middling activations for `pupils`, `clinic`, `research` this neuron might be firing for words related to employment

    oai_token-act-pair · gemini-2.0-flash

    Work

    np_max-act · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_6/width_16k/average_l0_95
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.6.ln2.hook_normalized
    Architecture
    jumprelu_transcoder
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     working
    -0.98
     Working
    -0.95
    Working
    -0.93
    working
    -0.92
     Work
    -0.91
     work
    -0.86
    Work
    -0.85
     WORKING
    -0.84
     worked
    -0.83
    worked
    -0.81
    POSITIVE LOGITS
    scriptcase
    0.59
    ollectionView
    0.59
    iettivo
    0.57
     BoxDecoration
    0.56
    APOLIS
    0.56
     consultato
    0.56
    /*++
    0.54
    atsun
    0.52
    ueger
    0.52
     autorytatywna
    0.52
    Activations Density 36.132%

    No Known Activations