© 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. 12216
    Prev
    Next
    INDEX
    Explanations

    conditional statements or assumptions regarding policies or rates

    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
    $MESS
    -0.17
    datagrid
    -0.15
    Depending
    -0.14
    953
    -0.14
    DBus
    -0.14
    WindowState
    -0.14
    -fontawesome
    -0.14
    pekt
    -0.14
    ÙİØ©
    -0.14
    @js
    -0.14
    POSITIVE LOGITS
     somehow
    0.26
     assumptions
    0.25
     assume
    0.25
     assumption
    0.24
     hypoth
    0.24
     perfect
    0.24
     typical
    0.23
     assumed
    0.22
    perfect
    0.22
     perfectly
    0.21
    Activations Density 0.029%

    No Known Activations