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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    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. 15898
    Prev
    Next
    INDEX
    Explanations

    discussions about assessments or comparisons of costs and coverage options in healthcare contexts

    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
     many
    -0.19
     Many
    -0.18
    _MANY
    -0.18
    many
    -0.17
    Many
    -0.16
     багаÑĤÑĮ
    -0.16
     muchos
    -0.16
     unless
    -0.15
     MANY
    -0.15
    unless
    -0.15
    POSITIVE LOGITS
     different
    0.70
     various
    0.61
    different
    0.56
     Different
    0.55
     Various
    0.54
    Different
    0.50
    ä¸įåIJĮçļĦ
    0.49
    ä¸įåIJĮ
    0.48
     diferentes
    0.48
     verschiedenen
    0.46
    Activations Density 0.035%

    No Known Activations