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. 31320
    Prev
    Next
    INDEX
    Explanations

    names of geographical places or locations

    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
    ocup
    -0.16
     concent
    -0.16
    ewn
    -0.16
    év
    -0.15
    ản
    -0.15
    ppo
    -0.14
    apia
    -0.14
    öm
    -0.14
    ÃŃlia
    -0.14
    áÄį
    -0.14
    POSITIVE LOGITS
    atchewan
    0.21
    illance
    0.18
    -serif
    0.17
    ENAME
    0.16
    ultz
    0.16
    ItemImage
    0.15
    ConfigurationException
    0.15
    PRINTF
    0.14
    عا
    0.14
    วรร
    0.14
    Activations Density 0.049%

    No Known Activations