© 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. 13481
    Prev
    Next
    INDEX
    Explanations

    instances of the word "talk" or its variations in the context of discussions or narratives

    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
    éº
    -0.17
    oran
    -0.15
    ilyn
    -0.15
    .fx
    -0.14
     Noah
    -0.14
     Alman
    -0.14
    ucas
    -0.14
    827
    -0.14
     voice
    -0.14
    vic
    -0.14
    POSITIVE LOGITS
    Talk
    0.32
     talk
    0.30
     Talk
    0.28
     about
    0.28
    talk
    0.27
    -talk
    0.27
     talked
    0.23
     talks
    0.21
     tentang
    0.20
    about
    0.20
    Activations Density 0.022%

    No Known Activations