© 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. 43159
    Prev
    Next
    INDEX
    Explanations

    mentions of conflicts of interest in decision-making 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
    reff
    -0.15
    spir
    -0.15
    :NSLocalizedString
    -0.15
    zimmer
    -0.15
    .Ordinal
    -0.14
    anager
    -0.14
    anded
    -0.14
    ÄĽj
    -0.14
    cript
    -0.13
    incare
    -0.13
    POSITIVE LOGITS
     conflicts
    0.63
     conflict
    0.62
     Conflict
    0.57
     conflic
    0.49
    Conflict
    0.48
     conflicting
    0.34
    -conf
    0.28
     confl
    0.27
     Conf
    0.27
     bias
    0.26
    Activations Density 0.050%

    No Known Activations