© 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. 44069
    Prev
    Next
    INDEX
    Explanations

    tweets that discuss accidents and related medical costs

    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
     pronounce
    -0.16
     Emails
    -0.16
     pronunciation
    -0.15
    opak
    -0.15
    .hwp
    -0.14
     emails
    -0.14
    åª
    -0.14
     interview
    -0.14
     Pron
    -0.14
    pron
    -0.14
    POSITIVE LOGITS
     caption
    0.28
     accompanied
    0.26
     wrote
    0.25
     accompanying
    0.25
     accompany
    0.24
     writing
    0.23
     alongside
    0.21
    caption
    0.20
     posts
    0.20
     accompanies
    0.19
    Activations Density 0.065%

    No Known Activations