Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Julius Han · Gated SAE for Llama3-8B-Instruct
    3. Llama3-8B-IT
    4. Residual Stream
    5. 25-RES-JH
    6. 23619
    Prev
    Next
    INDEX
    Explanations

    phrases that discuss health and dietary studies, particularly those related to sugar consumption and its effects

    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
    akra
    -0.17
    chied
    -0.16
    askell
    -0.15
    룬
    -0.14
     nowhere
    -0.14
    åĶĩ
    -0.14
    MAS
    -0.13
    ongan
    -0.13
     Suit
    -0.13
    sampling
    -0.13
    POSITIVE LOGITS
     quint
    0.20
     intake
    0.18
     exposure
    0.18
     leisure
    0.17
     exposures
    0.16
     REPLACE
    0.16
     daily
    0.16
     Exposure
    0.16
     Leisure
    0.16
     lag
    0.16
    Activations Density 0.048%

    No Known Activations