© 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. 21025
    Prev
    Next
    INDEX
    Explanations

    emotional expressions of grief and loss

    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
    agn
    -0.17
    Lon
    -0.14
     touching
    -0.14
    avanaugh
    -0.14
    entic
    -0.13
     cul
    -0.13
     compens
    -0.13
    ripper
    -0.13
    ilden
    -0.13
    ahn
    -0.13
    POSITIVE LOGITS
     shock
    0.23
     saddened
    0.22
     Shock
    0.22
     shocked
    0.21
    deep
    0.20
     shocks
    0.19
     sadness
    0.19
    æ·±
    0.18
     Deep
    0.18
     deep
    0.18
    Activations Density 0.046%

    No Known Activations