© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 41043
    Prev
    Next
    INDEX
    Explanations

    events related to a child's rescue and reunion with family

    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
    utin
    -0.17
     prep
    -0.15
     preparation
    -0.15
     Preparation
    -0.14
    /title
    -0.14
    ¶Į
    -0.14
    .bin
    -0.14
    ARSER
    -0.14
     plu
    -0.14
    udden
    -0.13
    POSITIVE LOGITS
     safe
    0.38
     safely
    0.36
    safe
    0.30
    à¸Ľà¸¥à¸Ńà¸Ķà¸ł
    0.28
     safety
    0.28
    Safe
    0.27
     escape
    0.27
     Safe
    0.26
     rescue
    0.26
     saf
    0.25
    Activations Density 0.078%

    No Known Activations