© 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. 14208
    Prev
    Next
    INDEX
    Explanations

    instances of the word "contradict."

    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
    aylight
    -0.16
    ê°Ŀ
    -0.15
     |-
    -0.15
    pard
    -0.15
    leyici
    -0.14
    iyat
    -0.14
    ainer
    -0.14
    andas
    -0.14
    owler
    -0.14
    oogle
    -0.14
    POSITIVE LOGITS
    -Sah
    0.15
    uli
    0.15
    iry
    0.14
    icts
    0.14
    .setDefault
    0.14
     ä½IJ
    0.14
    icted
    0.14
    ÑĢиз
    0.14
    à¹īà¸ķ
    0.14
    à¤Ĺर
    0.14
    Activations Density 0.001%

    No Known Activations