© 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. 43604
    Prev
    Next
    INDEX
    Explanations

    mentions of animal-related incidents, particularly involving horses and sexual misconduct

    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
    istrat
    -0.17
     dog
    -0.16
     Penguin
    -0.15
    æ¤ħ
    -0.15
     Dog
    -0.14
     cuck
    -0.14
    dog
    -0.14
     Jur
    -0.14
     Glover
    -0.14
    752
    -0.13
    POSITIVE LOGITS
     horses
    0.46
     horse
    0.45
     Horse
    0.40
     equ
    0.38
     mare
    0.37
     stall
    0.36
    horse
    0.34
     Equ
    0.34
     geld
    0.33
    馬
    0.30
    Activations Density 0.020%

    No Known Activations