© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 45012
    Prev
    Next
    INDEX
    Explanations

    phrases containing punctuation marks, particularly commas and periods

    oai_token-act-pair · gpt-4o-mini
    Triggered 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
    Negative Logits
    ektor
    -0.16
    ekim
    -0.15
    .TestTools
    -0.13
    ãģĹãģ¦ãĤĤ
    -0.13
    ÙĪÛĮØ´
    -0.13
    ÅĽ
    -0.13
    -↵
    -0.13
    ÐĤ
    -0.13
    íijľ
    -0.12
     queryInterface
    -0.12
    POSITIVE LOGITS
     according
    0.61
    according
    0.49
     According
    0.41
    According
    0.40
     selon
    0.33
    ccording
    0.31
     sources
    0.31
     según
    0.30
     accordance
    0.28
     said
    0.27
    Activations Density 0.306%

    No Known Activations