© 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. 63637
    Prev
    Next
    INDEX
    Explanations

    government and policy-related discussions, particularly in the context of waste management and regulations

    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
     honored
    -0.18
    Center
    -0.18
     honors
    -0.18
     favored
    -0.17
    center
    -0.16
     Presidential
    -0.16
     Lafayette
    -0.16
     Counseling
    -0.16
     honor
    -0.16
     harbor
    -0.16
    POSITIVE LOGITS
     RCMP
    0.36
     Winnipeg
    0.36
    .ca
    0.35
     Saskatchewan
    0.34
     Ontario
    0.32
     Manitoba
    0.32
     Edmonton
    0.32
    Toronto
    0.32
     Ottawa
    0.31
     Canadians
    0.31
    Activations Density 0.050%

    No Known Activations