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. 23551
    Prev
    Next
    INDEX
    Explanations

    topics related to music and lyrics, as well as issues around regulations and consumer protections

    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
    711
    -0.15
    ike
    -0.14
    851
    -0.14
    æĹ©
    -0.13
    uchen
    -0.13
    Aspect
    -0.13
     Roe
    -0.13
     Ear
    -0.13
    nech
    -0.13
     transformation
    -0.13
    POSITIVE LOGITS
     added
    0.24
    -old
    0.23
     arrivals
    0.22
    (old
    0.21
    sworth
    0.21
     introduced
    0.21
    /new
    0.21
    added
    0.20
     Added
    0.20
    -added
    0.20
    Activations Density 0.086%

    No Known Activations