© 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 EvalsExportsinterp-engineNEWAPI 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. 44608
    Prev
    Next
    INDEX
    Explanations

    references to threats and protection for individuals facing danger

    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
    gorit
    -0.15
     inflict
    -0.14
    iet
    -0.14
     Ø®ÙĦ
    -0.14
    uyết
    -0.13
    Æ¡
    -0.13
    isch
    -0.13
     جÙħعÛĮت
    -0.13
     inflicted
    -0.13
     fırsat
    -0.13
    POSITIVE LOGITS
     threats
    0.35
     intimidation
    0.31
     intimid
    0.30
     intimidated
    0.28
     threatened
    0.28
     Threat
    0.28
     threat
    0.26
     intimidate
    0.26
    threat
    0.25
     repr
    0.25
    Activations Density 0.070%

    No Known Activations