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. 96
    Prev
    Next
    INDEX
    Explanations

    names and titles of government officials or senators

    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
    elyn
    -0.18
     Lange
    -0.16
    uple
    -0.15
    碼
    -0.15
    åĺ
    -0.14
     repell
    -0.14
    еÑĪ
    -0.14
    oldown
    -0.14
    aight
    -0.14
     Butter
    -0.14
    POSITIVE LOGITS
     senator
    0.24
     senators
    0.23
     senate
    0.23
     Senate
    0.23
     Senator
    0.22
     Senators
    0.22
    Senator
    0.20
    Sen
    0.19
    Senate
    0.19
    .sensor
    0.18
    Activations Density 0.018%

    No Known Activations