© 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. 12140
    Prev
    Next
    INDEX
    Explanations

    concepts related to cancer treatment and clinical trials

    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
     preventative
    -0.17
    orer
    -0.16
     Til
    -0.14
     developmental
    -0.14
    tul
    -0.13
    ycz
    -0.13
    612
    -0.13
    prite
    -0.13
    Ñįн
    -0.13
    siz
    -0.13
    POSITIVE LOGITS
     treatment
    0.86
     treatments
    0.77
     Treatment
    0.72
    Treatment
    0.61
     Treat
    0.57
     treat
    0.56
     treating
    0.54
    reatment
    0.52
     behand
    0.51
     tratamiento
    0.51
    Activations Density 0.071%

    No Known Activations