© 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. 13175
    Prev
    Next
    INDEX
    Explanations

    indicators of emphasis or significance in a text

    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
    lessly
    -0.19
    ëĭĪëĭ¤
    -0.18
    ively
    -0.17
    Vectorizer
    -0.15
    attles
    -0.15
    imately
    -0.15
    ìłģìľ¼ë¡ľ
    -0.15
    ÅĻi
    -0.14
    æĮģãģ¡
    -0.14
    zelf
    -0.13
    POSITIVE LOGITS
    sembl
    0.15
    istance
    0.15
    ystore
    0.14
     Kraj
    0.14
    dden
    0.14
    ness
    0.14
     Jacobs
    0.14
    ‘s
    0.14
    ÑİваннÑı
    0.13
    алÑĭ
    0.13
    Activations Density 0.001%

    No Known Activations