© 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. 21206
    Prev
    Next
    INDEX
    Explanations

    references to supernatural or demonic entities in the context of moral dilemmas

    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
    ault
    -0.17
    ignet
    -0.16
    rese
    -0.15
     Creed
    -0.15
    estr
    -0.15
    ippi
    -0.14
    ãĥ³ãĥĸ
    -0.14
    æĬŀ
    -0.14
    ลาà¸Ķ
    -0.14
     Warrior
    -0.14
    POSITIVE LOGITS
     devil
    0.56
     sat
    0.53
     Sat
    0.50
     Devil
    0.50
    devil
    0.45
     demon
    0.45
     Satan
    0.43
     dev
    0.43
     Devils
    0.41
     Demon
    0.40
    Activations Density 0.032%

    No Known Activations