© 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. 13786
    Prev
    Next
    INDEX
    Explanations

    phrases related to claims, comments, or attempts in reporting and discussing information

    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
    mvc
    -0.14
    umlu
    -0.14
    ÄĽr
    -0.14
    rias
    -0.13
    enze
    -0.13
    wiÄħ
    -0.13
     muschi
    -0.13
    оÑģÑĥд
    -0.13
    elijk
    -0.13
    à¹Ģà¸Ł
    -0.13
    POSITIVE LOGITS
     made
    1.17
    made
    1.02
     Made
    0.96
    Made
    0.93
     MADE
    0.84
    -made
    0.84
    emade
    0.59
     make
    0.58
     gemacht
    0.52
     Make
    0.51
    Activations Density 0.048%

    No Known Activations