© 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. 63502
    Prev
    Next
    INDEX
    Explanations

    instances of the word "are" and its variations in various contexts related to societal issues and injustices

    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
    elve
    -0.14
    .ToDateTime
    -0.14
    ablish
    -0.14
    apore
    -0.14
    idis
    -0.14
    imest
    -0.14
     VÅ¡ech
    -0.13
    amanho
    -0.13
    cloak
    -0.13
    .gridColumn
    -0.13
    POSITIVE LOGITS
     among
    0.24
     often
    0.23
     themselves
    0.22
     amongst
    0.21
     more
    0.20
     flock
    0.19
     not
    0.19
     targets
    0.18
     taught
    0.18
     generally
    0.18
    Activations Density 0.260%

    No Known Activations