© 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. 15749
    Prev
    Next
    INDEX
    Explanations

    numerical relationships and ratios within health and financial 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
    pla
    -0.16
    ogle
    -0.16
     Frank
    -0.15
    ainter
    -0.14
    .Max
    -0.14
    éĺ¶
    -0.14
    /testing
    -0.14
    _effects
    -0.14
    åłĤ
    -0.14
    /hooks
    -0.14
    POSITIVE LOGITS
     ratio
    0.89
     ratios
    0.78
     Ratio
    0.77
    ratio
    0.70
    Ratio
    0.67
    _ratio
    0.62
     Rat
    0.49
    _RATIO
    0.47
     ÑģооÑĤ
    0.41
    -to
    0.39
    Activations Density 0.038%

    No Known Activations