© 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. 12515
    Prev
    Next
    INDEX
    Explanations

    phrases related to tension and uncertainty in a conflict or competitive context

    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
     AFTER
    -0.15
    719
    -0.15
    ijken
    -0.15
    èįĴ
    -0.15
     gradually
    -0.14
    618
    -0.14
    ')?>
    -0.13
     temporarily
    -0.13
    ниÑĨÑĮ
    -0.13
     slowly
    -0.13
    POSITIVE LOGITS
     anytime
    0.57
     any
    0.41
     ANY
    0.36
    any
    0.33
     Any
    0.32
    Any
    0.31
    -any
    0.31
     anywhere
    0.30
    ä»»ä½ķ
    0.30
     cualquier
    0.30
    Activations Density 0.048%

    No Known Activations