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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 41681
    Prev
    Next
    INDEX
    Explanations

    phrases related to caution or consideration before taking action

    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
    GetCurrent
    -0.15
    laden
    -0.14
    oser
    -0.14
    rove
    -0.14
    _defs
    -0.13
    aida
    -0.13
    imd
    -0.13
    our
    -0.13
    Finding
    -0.13
     Finding
    -0.13
    POSITIVE LOGITS
    /if
    0.17
     they
    0.17
    uctor
    0.17
    hti
    0.17
     you
    0.17
     making
    0.16
     proceeding
    0.15
     we
    0.15
     sebou
    0.15
     deciding
    0.15
    Activations Density 0.057%

    No Known Activations