© 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. 21-GEMMASCOPE-TRANSCODER-16K
    7. 4667
    Prev
    Next
    INDEX
    Explanations

    This neuron appears to identify formal or academic text, potentially recognizing elements like citations, theorems, or structured arguments.

    oai_token-act-pair · gemini-2.0-flash

    tokens in different languages

    np_token-act-pair-logits · gpt-4o-mini

    say "we" or "I"

    np_max-act-logits · gemini-2.0-flash
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    google/gemma-scope-2b-pt-transcoders/layer_21/width_16k/average_l0_13
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.21.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
     there
    -0.95
    there
    -0.74
     they
    -0.72
    IntoConstraints
    -0.70
     this
    -0.69
     the
    -0.65
     THERE
    -0.63
     these
    -0.63
     those
    -0.62
    writeFieldEnd
    -0.61
    POSITIVE LOGITS
     nous
    0.53
     us
    0.52
     تعدى
    0.51
     Nous
    0.50
    umuz
    0.50
    ūsų
    0.49
     nakalista
    0.49
    RestTemplate
    0.48
     parceiros
    0.47
     مشين
    0.46
    Activations Density 1.475%

    No Known Activations