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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 28085
    Prev
    Next
    INDEX
    Explanations

    phrases indicating risk-taking behavior or the act of seizing opportunities

    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
    _UTF
    -0.16
    jin
    -0.15
    afil
    -0.15
    (Bit
    -0.14
    aad
    -0.14
     Král
    -0.14
     argent
    -0.14
    icker
    -0.14
    licting
    -0.14
    amma
    -0.14
    POSITIVE LOGITS
     chances
    0.59
     chance
    0.58
     risks
    0.47
     Chance
    0.45
     risk
    0.45
    Chance
    0.43
    risk
    0.34
     Risk
    0.33
    _chance
    0.33
    Risk
    0.33
    Activations Density 0.019%

    No Known Activations