© 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. 44052
    Prev
    Next
    INDEX
    Explanations

    references to personal commitments and scheduling

    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
     Pic
    -0.14
    ChangeEvent
    -0.14
     summers
    -0.14
     Dorm
    -0.14
    زة
    -0.14
    anik
    -0.14
    qui
    -0.13
    Pic
    -0.13
    SEL
    -0.13
     struggle
    -0.13
    POSITIVE LOGITS
     commitments
    0.36
     obligations
    0.34
     responsibilities
    0.31
     duties
    0.28
     appointments
    0.28
     err
    0.27
     appointment
    0.27
     work
    0.26
     important
    0.25
     commitment
    0.24
    Activations Density 0.092%

    No Known Activations