© 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. 42481
    Prev
    Next
    INDEX
    Explanations

    phrases related to tracking and reporting physical restraints in schools

    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
     Darling
    -0.17
    289
    -0.16
     Luc
    -0.16
     Ros
    -0.16
     ros
    -0.16
     Rom
    -0.15
    .scalablytyped
    -0.15
    Ros
    -0.15
    ros
    -0.15
    rates
    -0.13
    POSITIVE LOGITS
     restriction
    1.03
     restrictions
    0.93
    Restr
    0.89
     Restr
    0.87
    restrict
    0.84
    restriction
    0.84
     Restrictions
    0.84
     restricted
    0.83
     restrict
    0.83
     restr
    0.77
    Activations Density 0.011%

    No Known Activations