© 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. 8923
    Prev
    Next
    INDEX
    Explanations

    conditions and restrictions regarding ownership and use of products, particularly related to safety and compliance

    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
     unl
    -0.15
    unner
    -0.14
    antal
    -0.14
    одо
    -0.13
    лÑıн
    -0.13
    enga
    -0.13
    ebo
    -0.13
    /fs
    -0.13
    anya
    -0.13
    jee
    -0.13
    POSITIVE LOGITS
     approved
    0.48
    -approved
    0.42
     official
    0.41
     authorized
    0.41
    approved
    0.39
     officially
    0.35
     legitimate
    0.34
     legit
    0.33
    Approved
    0.33
     Approved
    0.33
    Activations Density 0.052%

    No Known Activations