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. 23594
    Prev
    Next
    INDEX
    Explanations

    phrases indicating emergency response actions or calls

    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
    ReadStream
    -0.17
    ÙħÛĮÙĦ
    -0.14
    ptal
    -0.14
     Escape
    -0.14
    reeze
    -0.14
     Himself
    -0.14
    ainless
    -0.13
    Escape
    -0.13
     himself
    -0.13
    afil
    -0.13
    POSITIVE LOGITS
     response
    0.27
     investigate
    0.26
     respond
    0.25
     responds
    0.23
     Response
    0.22
     investigates
    0.22
     responded
    0.22
     scene
    0.21
    response
    0.20
     responding
    0.20
    Activations Density 0.035%

    No Known Activations