© 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. 42841
    Prev
    Next
    INDEX
    Explanations

    instances of the word "promised" and its variations, indicating a focus on promises or commitments made by individuals or entities

    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
    zin
    -0.16
    remen
    -0.15
    yi
    -0.15
    alach
    -0.14
    ac
    -0.14
     requirement
    -0.14
    /of
    -0.14
    à¸ļà¸ģ
    -0.14
    397
    -0.14
    lam
    -0.13
    POSITIVE LOGITS
     to
    0.37
    to
    0.21
     να
    0.18
    	to
    0.17
     that
    0.17
    that
    0.16
     tp
    0.15
    ascus
    0.15
    yonel
    0.15
    ään
    0.15
    Activations Density 0.041%

    No Known Activations