© 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. Julius Han · Gated SAE for Llama3-8B-Instruct
    5. Llama3-8B-IT
    7. Residual Stream
    9. 25-RES-JH
    11. 12105
    Prev
    Next
    INDEX
    Explanations

    references to recent publications and special editions in media

    oai_token-act-pair · gpt-4o-mini
    Triggered 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
    Negative Logits
     cab
    -0.13
    enge
    -0.13
     Elf
    -0.13
    Elf
    -0.12
    olas
    -0.12
    IDEOS
    -0.12
    quoi
    -0.12
    olini
    -0.12
    regar
    -0.12
     aks
    -0.12
    POSITIVE LOGITS
     issue
    1.29
     Issue
    1.13
    issue
    1.07
    Issue
    1.02
     issues
    1.00
    -issue
    0.93
    .issue
    0.91
    _issue
    0.89
     ISSUE
    0.89
     Issues
    0.89
    Activations Density 0.041%

    No Known Activations