© 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 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. 14285
    Prev
    Next
    INDEX
    Explanations

    terms related to dietary fat and its impact on health

    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
    rani
    -0.18
    inaire
    -0.17
    pmat
    -0.16
    oded
    -0.15
    enticate
    -0.15
    innie
    -0.15
    aley
    -0.15
     Tou
    -0.14
    .dds
    -0.14
     Casc
    -0.14
    POSITIVE LOGITS
    igue
    0.21
     soluble
    0.21
     content
    0.21
     glob
    0.20
    Content
    0.20
    /o
    0.19
     Content
    0.19
    ting
    0.18
    iness
    0.17
     acids
    0.17
    Activations Density 0.020%

    No Known Activations