© 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. 43315
    Prev
    Next
    INDEX
    Explanations

    key terms related to choice and evaluation of options

    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
    Ïģιά
    -0.16
    969
    -0.13
    459
    -0.13
    forman
    -0.13
    abar
    -0.12
    oltip
    -0.12
    tml
    -0.12
    ergus
    -0.12
    porno
    -0.12
    pert
    -0.12
    POSITIVE LOGITS
     over
    0.91
     sobre
    0.58
    Over
    0.51
     över
    0.51
    over
    0.51
     OVER
    0.50
    _over
    0.50
     Over
    0.49
    è¿ĩ
    0.46
     über
    0.45
    Activations Density 0.041%

    No Known Activations