© 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-2-2B
    5. 25-GEMMASCOPE-TRANSCODER-16K
    7. 5217
    Prev
    Next
    INDEX
    Explanations

    varied. The most common activation pattern is triggered by the words "be" and "was" in the context of medical and environmental texts. It also activates for numbers and medical terminology. This suggests that the neuron is identifying a wide range of potentially important topics

    oai_token-act-pair · gemini-2.0-flash

    problems being addressed

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_25/width_16k/average_l0_41
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.25.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
     reparar
    -0.65
     repair
    -0.63
    repair
    -0.58
     repairing
    -0.58
    Repair
    -0.57
    dymyr
    -0.57
    recover
    -0.57
    recovery
    -0.57
     fix
    -0.56
     Repair
    -0.56
    POSITIVE LOGITS
    angu
    0.51
     jsPsych
    0.51
    DoubleQuotes
    0.50
    lister
    0.49
     Lijst
    0.49
     Drawing
    0.48
    aritions
    0.47
     expressing
    0.47
     drawing
    0.47
    Referencer
    0.47
    Activations Density 8.887%

    No Known Activations