© 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. 45062
    Prev
    Next
    INDEX
    Explanations

    instances of aggression in various contexts

    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
    ergus
    -0.15
    elage
    -0.15
    ariat
    -0.14
    otope
    -0.14
    oden
    -0.14
    erval
    -0.14
    ekl
    -0.14
    ican
    -0.14
    uze
    -0.13
    oples
    -0.13
    POSITIVE LOGITS
    ging
    0.35
    ged
    0.30
    gy
    0.26
    gers
    0.26
    gle
    0.21
    GING
    0.21
    ressive
    0.20
    ression
    0.19
    mented
    0.19
    gings
    0.19
    Activations Density 0.001%

    No Known Activations