© 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. 44005
    Prev
    Next
    INDEX
    Explanations

    the word "the" in various contexts

    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
    //{{
    -0.18
    zug
    -0.14
    enburg
    -0.14
     eiusmod
    -0.14
    ulace
    -0.13
    emachine
    -0.13
    лаÑĩ
    -0.13
     tablesp
    -0.13
    ุ
    -0.13
    =}
    -0.13
    POSITIVE LOGITS
     meantime
    0.21
     morning
    0.21
     midst
    0.20
     hopes
    0.20
     beginning
    0.20
     middle
    0.20
     sense
    0.19
     immortal
    0.19
     meanwhile
    0.18
     shower
    0.18
    Activations Density 0.161%

    No Known Activations