© 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. 44571
    Prev
    Next
    INDEX
    Explanations

    occurrences of the verb "is" and its variations, indicating a focus on linking or stating existence

    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
    à¤ľà¤°
    -0.15
    pone
    -0.15
    OLLOW
    -0.15
     previously
    -0.14
     olmaz
    -0.13
     Kaf
    -0.13
    orgot
    -0.13
    /archive
    -0.13
     unearth
    -0.13
    493
    -0.13
    POSITIVE LOGITS
     currently
    0.41
     now
    0.40
    currently
    0.36
     being
    0.34
    now
    0.29
     presently
    0.28
    being
    0.28
     Currently
    0.27
     still
    0.27
    Currently
    0.27
    Activations Density 0.262%

    No Known Activations