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

    Neuronpedia

    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. 55164
    Prev
    Next
    INDEX
    Explanations

    phrases related to confusion and chaotic situations

    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
    ź
    -0.15
    ruk
    -0.15
    longleftrightarrow
    -0.15
    åł¡
    -0.14
    à¸ļาย
    -0.14
    Ñİк
    -0.13
    rière
    -0.13
    زش
    -0.13
    uka
    -0.13
    زÛĮ
    -0.13
    POSITIVE LOGITS
     trying
    0.77
     try
    0.64
    trying
    0.61
     Trying
    0.60
     tries
    0.58
    Trying
    0.55
    try
    0.54
     Try
    0.53
     TRY
    0.53
    Try
    0.49
    Activations Density 0.072%

    No Known Activations