© 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 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. 14790
    Prev
    Next
    INDEX
    Explanations

    references to additional information and resources

    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
    lech
    -0.15
     Ih
    -0.14
    ÑģÑĥÑĤ
    -0.14
    ÙĪØ·
    -0.13
    çĿĢ
    -0.13
    BTN
    -0.13
     èĴ
    -0.13
    _REUSE
    -0.13
    ếp
    -0.12
    xdf
    -0.12
    POSITIVE LOGITS
     about
    0.38
    about
    0.29
     on
    0.27
     tentang
    0.25
     vá»ģ
    0.22
    åħ³äºİ
    0.21
     About
    0.21
     ABOUT
    0.21
    About
    0.20
     عÙĨÙĩ
    0.18
    Activations Density 0.070%

    No Known Activations