© 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. 12158
    Prev
    Next
    INDEX
    Explanations

    conjunctions and pronouns indicating contrasting situations or conditions

    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
     therefore
    -0.24
     indeed
    -0.23
     Even
    -0.20
     zwar
    -0.19
     even
    -0.19
     donc
    -0.19
     Therefore
    -0.18
    tplib
    -0.18
     accordingly
    -0.18
     tháºŃm
    -0.17
    POSITIVE LOGITS
     soon
    0.33
    soon
    0.27
     quickly
    0.22
     Soon
    0.22
    Soon
    0.22
     later
    0.20
    å¾Įãģ«
    0.19
     ultimately
    0.19
    ler
    0.19
     afterwards
    0.18
    Activations Density 0.067%

    No Known Activations