© 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. 63602
    Prev
    Next
    INDEX
    Explanations

    instances of the word "switch" or its derivatives, indicating changes or transitions in a situation or context

    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
    ture
    -0.16
    ï½¥
    -0.15
    ilet
    -0.14
    .environment
    -0.14
    /site
    -0.14
    ajo
    -0.14
    лÑİд
    -0.13
    γκο
    -0.13
    adoo
    -0.13
    OfType
    -0.13
    POSITIVE LOGITS
     gears
    0.30
     sides
    0.26
     roles
    0.24
     focus
    0.22
     sexes
    0.20
     allegiance
    0.20
     alleg
    0.20
     lanes
    0.20
     tack
    0.19
     directions
    0.19
    Activations Density 0.052%

    No Known Activations