Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlog/PodcastSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlog/PodcastGitHubSlackTwitterContact
    1. Home
    2. Julius Han · Gated SAE for Llama3-8B-Instruct
    3. Llama3-8B-IT
    4. Residual Stream
    5. 25-RES-JH
    6. 23617
    Prev
    Next
    INDEX
    Explanations

    references to singular entities or unique situations

    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
     or
    -0.18
     such
    -0.18
    amage
    -0.17
     either
    -0.17
     ones
    -0.16
     various
    -0.16
     next
    -0.16
    инок
    -0.16
     alike
    -0.15
    °E
    -0.15
    POSITIVE LOGITS
     each
    0.19
    Each
    0.18
    OTHER
    0.18
     entire
    0.18
     Each
    0.18
     instance
    0.17
    others
    0.17
    other
    0.17
    each
    0.17
     EACH
    0.17
    Activations Density 0.011%

    No Known Activations