© 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. 21756
    Prev
    Next
    INDEX
    Explanations

    emotional and expressive language related to positive experiences

    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
    adero
    -0.14
    ëıĦë¡ľ
    -0.14
    peg
    -0.14
    ãĥªãĥ¼ãĤº
    -0.14
    onth
    -0.13
    ain
    -0.13
    Appending
    -0.12
    ocus
    -0.12
    .userData
    -0.12
    ndx
    -0.12
    POSITIVE LOGITS
     seeing
    0.39
     to
    0.39
     hearing
    0.33
     having
    0.32
     knowing
    0.32
    seeing
    0.29
    having
    0.26
     να
    0.26
     being
    0.25
     Seeing
    0.25
    Activations Density 0.082%

    No Known Activations