© 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. 13292
    Prev
    Next
    INDEX
    Explanations

    numerical statistics related to performance metrics

    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
    lander
    -0.16
    libft
    -0.16
     pads
    -0.14
     pent
    -0.14
    oir
    -0.14
    oa
    -0.14
    ä¿Ĺ
    -0.14
    OfSize
    -0.14
    67
    -0.14
    VOID
    -0.14
    POSITIVE LOGITS
    120
    0.30
    80
    0.30
    160
    0.29
    40
    0.28
    70
    0.28
    140
    0.28
    60
    0.28
    520
    0.27
    560
    0.26
    90
    0.26
    Activations Density 0.028%

    No Known Activations