© 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. 12438
    Prev
    Next
    INDEX
    Explanations

    this neuron identifies cause and effect relationships and conditional statements.

    oai_token-act-pair · gemini-2.0-flash

    programming or technical terms

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

    code/technical text

    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
     ones
    -2.44
     them
    -2.34
     those
    -2.09
     They
    -2.08
     they
    -2.06
    They
    -2.05
    those
    -1.91
    them
    -1.88
    它们
    -1.87
    they
    -1.81
    POSITIVE LOGITS
     resourceCulture
    0.65
    }>{
    0.56
    RegressionTest
    0.54
    újo
    0.53
     she
    0.51
    utique
    0.50
    \{\\
    0.50
    getline
    0.50
    žený
    0.50
     berdua
    0.49
    Activations Density 5.395%

    No Known Activations