© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    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. 22898
    Prev
    Next
    INDEX
    Explanations

    instances of the word "down" and related phrases

    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
    abei
    -0.17
    igon
    -0.17
     up
    -0.17
    phants
    -0.16
    olean
    -0.16
    jezd
    -0.15
    Ä±ÅŁÄ±k
    -0.15
     dán
    -0.15
    gesi
    -0.14
     downtown
    -0.14
    POSITIVE LOGITS
    graded
    0.27
    wards
    0.26
    ed
    0.26
    /up
    0.25
    grades
    0.23
    played
    0.22
    grading
    0.21
    ward
    0.20
    stairs
    0.19
    pour
    0.19
    Activations Density 0.063%

    No Known Activations