© 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. 14944
    Prev
    Next
    INDEX
    Explanations

    references to competitive sports and racing

    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
    typeof
    -0.15
    lient
    -0.15
    .SDK
    -0.14
    ÑıÑī
    -0.14
    pell
    -0.14
     звиÑĩай
    -0.14
    URRED
    -0.13
    ropol
    -0.13
    onte
    -0.13
    ëŀijìĬ¤
    -0.13
    POSITIVE LOGITS
     having
    1.13
    having
    0.98
     Having
    0.96
    Having
    0.90
     ayant
    0.64
     avoir
    0.59
     hav
    0.52
     haber
    0.49
     имеÑĤÑĮ
    0.38
     memiliki
    0.38
    Activations Density 0.023%

    No Known Activations