© 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. 12713
    Prev
    Next
    INDEX
    Explanations

    phrases referring to specific elements or concepts

    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
    (æ°´
    -0.15
    (æľĪ
    -0.15
    ÑĸÑģ
    -0.14
     legislation
    -0.13
    (æĹ¥
    -0.13
    iveau
    -0.13
    ìłģ
    -0.13
    ATUS
    -0.13
    Č
    -0.13
    (åľŁ
    -0.13
    POSITIVE LOGITS
     same
    0.28
    curity
    0.27
     latter
    0.27
     kinds
    0.26
     particular
    0.25
    -ÑĤо
    0.24
    cond
    0.22
    quence
    0.22
    zelf
    0.21
     respective
    0.20
    Activations Density 1.108%

    No Known Activations