© 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. 22051
    Prev
    Next
    INDEX
    Explanations

    phrases indicating uncertainty or questioning capability

    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
    elerik
    -0.14
    xito
    -0.14
     Ðļаб
    -0.14
     Lorem
    -0.14
    razier
    -0.14
    $MESS
    -0.14
    ught
    -0.13
    ]={↵
    -0.13
    ÙĪز
    -0.13
    "./
    -0.13
    POSITIVE LOGITS
     cannot
    0.54
    cannot
    0.42
     cant
    0.41
     cann
    0.41
     Cannot
    0.41
     can
    0.41
    Cannot
    0.36
     unable
    0.36
     CAN
    0.34
     Cant
    0.29
    Activations Density 0.111%

    No Known Activations