© 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. 22957
    Prev
    Next
    INDEX
    Explanations

    modals that express future intentions or possibilities

    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
    itzer
    -0.18
    udas
    -0.15
     NOT
    -0.14
    arga
    -0.14
    aire
    -0.13
    rou
    -0.13
     slightest
    -0.13
    ä¸įäºĨ
    -0.13
    ithe
    -0.13
    dda
    -0.13
    POSITIVE LOGITS
     always
    0.75
    always
    0.66
     Always
    0.63
    Always
    0.62
     sempre
    0.61
     toujours
    0.59
     siempre
    0.58
     ALWAYS
    0.53
     вÑģегда
    0.49
     altijd
    0.47
    Activations Density 0.077%

    No Known Activations