© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Julius Han · Gated SAE for Llama3-8B-Instruct
    5. Llama3-8B-IT
    7. Residual Stream
    9. 25-RES-JH
    11. 44734
    Prev
    Next
    INDEX
    Explanations

    instances of the word "the" and related variants

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    Juliushanhanhan/llama-3-8b-it-res/blocks.25.hook_resid_post
    Features
    65,536
    Data Type
    float32
    Hook Name
    blocks.25.hook_resid_post
    Hook Layer
    25
    Architecture
    gated
    Context Size
    1,024
    Dataset
    Juliushanhanhan/openwebtext-1b-llama3-tokenized-cxt-1024
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     MetroFramework
    -0.15
     duy
    -0.14
     dahi
    -0.14
    aeda
    -0.13
     ../../../
    -0.13
    lest
    -0.13
    -next
    -0.13
    adÃŃ
    -0.13
    drv
    -0.13
    oran
    -0.13
    POSITIVE LOGITS
     latter
    0.29
     smaller
    0.27
     first
    0.27
     more
    0.27
     older
    0.26
     younger
    0.25
     earlier
    0.24
     newer
    0.23
     shorter
    0.23
     larger
    0.23
    Activations Density 0.142%

    No Known Activations