Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Gemma-2-2B
    3. 23-GEMMASCOPE-TRANSCODER-16K
    4. 12097
    Prev
    Next
    INDEX
    Explanations

    This neuron appears to be looking for a mix of things including URLs with youtube, the word "moreover," code snippets, and potentially other markers of technical writing or discussion.

    oai_token-act-pair · gemini-2.0-flash

    code snippets

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_23/width_16k/average_l0_25
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.23.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
     
    -0.42
    <eos>
    -0.42
    1
    -0.37
    <unused61>
    -0.37
    <unused63>
    -0.37
    2
    -0.35
    <unused62>
    -0.33
    &
    -0.33
    4
    -0.31
    3
    -0.31
    POSITIVE LOGITS
     Theſe
    2.22
     Efq
    2.09
     Monfieur
    2.05
     myſelf
    1.95
     itſelf
    1.89
     Jefus
    1.85
     ―――――
    1.79
    parsedMessage
    1.78
    <unused47>
    1.76
    <unused99>
    1.76
    Activations Density 3.210%

    No Known Activations