© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Julius Han · Gated SAE for Llama3-8B-Instruct
    5. Llama3-8B-IT
    7. Residual Stream
    9. 25-RES-JH
    11. 32979
    Prev
    Next
    INDEX
    Explanations

    instances of notable names or people mentioned in discussions

    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
     nicht
    -0.30
     tidak
    -0.30
     não
    -0.28
     ikke
    -0.28
     không
    -0.28
     not
    -0.28
     не
    -0.27
    ä¸įæĺ¯
    -0.26
     नह
    -0.24
     niet
    -0.24
    POSITIVE LOGITS
     instead
    0.33
     Instead
    0.30
     opt
    0.29
    Instead
    0.29
    instead
    0.28
     simply
    0.27
     merely
    0.27
     opting
    0.24
     prefer
    0.23
     opted
    0.22
    Activations Density 0.038%

    No Known Activations