© 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. 13091
    Prev
    Next
    INDEX
    Explanations

    phrases related to exacerbation or worsening conditions, particularly in social, economic, or environmental 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
    ahi
    -0.16
    pit
    -0.15
    ivar
    -0.15
    à¹īà¸Ńà¸ĩ
    -0.15
    cott
    -0.14
    ilk
    -0.14
    WidgetItem
    -0.13
    ocker
    -0.13
    itory
    -0.13
    zp
    -0.13
    POSITIVE LOGITS
     already
    0.38
     Already
    0.35
    already
    0.34
     existing
    0.33
    Already
    0.32
    _already
    0.28
     Existing
    0.28
     matters
    0.27
     Matters
    0.26
    existing
    0.25
    Activations Density 0.025%

    No Known Activations