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. 23235
    Prev
    Next
    INDEX
    Explanations

    mentions of specific cities, particularly San Francisco and Los Angeles

    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
    loyd
    -0.16
    lies
    -0.15
    cky
    -0.14
     invisible
    -0.14
     Blanch
    -0.14
    etler
    -0.14
    utc
    -0.13
    wright
    -0.13
    ierge
    -0.13
     inn
    -0.13
    POSITIVE LOGITS
    neapolis
    0.25
    adelphia
    0.24
     Moines
    0.24
    åı¤å±ĭ
    0.22
    sterdam
    0.22
    furt
    0.22
     Hague
    0.21
    olulu
    0.21
     Aires
    0.21
    zburg
    0.20
    Activations Density 0.080%

    No Known Activations