© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW Community BlogPrivacy & TermsContact
    1. Home
    3. Google DeepMind · Exploring Gemma 2 with Gemma Scope
    5. Gemma-2-9B
    7. Residual Stream - 16k
    9. 31-GEMMASCOPE-RES-16K
    11. 2957
    Prev
    Next
    INDEX
    Explanations

    technical or mathematical notations, including variables and their properties

    oai_token-act-pair · gpt-4o-mini
    Triggered by @bot

    The neuron activates for lowercase letter 'd' in technical or code-related contexts.

    oai_token-act-pair · claude-3-7-sonnet-20250219
    Triggered by @neilrathi

    hexadecimal notation and numeric codes (in the format 0x followed by alphanumeric characters).

    oai_token-act-pair · claude-4-5-haiku
    Triggered by @emiglarou
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GEMMA-2-9B @ 31-gemmascope-res-16k
    Configuration
    google/gemma-scope-9b-pt-res/layer_31/width_16k/average_l0_114
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    16,384
    Data Type
    float32
    Hook Name
    blocks.31.hook_resid_post
    Hook Layer
    31
    Architecture
    jumprelu
    Context Size
    1,024
    Dataset
    monology/pile-uncopyrighted
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Negative Logits
    पया
    -0.54
    \{\\
    -0.54
    ContentAlignment
    -0.52
     antemano
    -0.50
     ListTile
    -0.50
     libremente
    -0.47
     assurent
    -0.47
    ItemBackground
    -0.47
    iers
    -0.47
     queſta
    -0.46
    POSITIVE LOGITS
     rad
    0.39
    l
    0.38
    ChromeDriver
    0.37
     Fund
    0.35
    getFont
    0.35
     FontAwesome
    0.35
     Font
    0.35
    žad
    0.34
     userDao
    0.34
     fund
    0.33
    Activations Density 0.568%

    No Known Activations