© 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. 40702
    Prev
    Next
    INDEX
    Explanations

    depictions of characters and their traits in 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
    pong
    -0.17
    Alive
    -0.15
    yonel
    -0.15
    ÑģоÑĤ
    -0.15
    ortal
    -0.14
    rana
    -0.14
    aju
    -0.14
    okit
    -0.14
    alive
    -0.14
    Porno
    -0.14
    POSITIVE LOGITS
     hero
    0.35
     heroic
    0.28
     Hero
    0.28
    -hero
    0.26
    hero
    0.25
     victim
    0.24
     heroes
    0.24
     figure
    0.22
    Hero
    0.22
     martyr
    0.21
    Activations Density 0.074%

    No Known Activations