© 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. 25778
    Prev
    Next
    INDEX
    Explanations

    descriptors of human facial expressions and emotions, particularly related to aging and struggle

    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
    929
    -0.16
     Rust
    -0.14
     Confeder
    -0.14
     tail
    -0.14
    ean
    -0.14
     retro
    -0.14
    ipop
    -0.14
    inem
    -0.14
     body
    -0.13
    xac
    -0.13
    POSITIVE LOGITS
     pale
    0.21
     bath
    0.20
     streak
    0.20
     Pall
    0.18
     pall
    0.18
    chalk
    0.18
     covered
    0.18
     puck
    0.17
     ga
    0.17
    flush
    0.17
    Activations Density 0.023%

    No Known Activations