© 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. 45307
    Prev
    Next
    INDEX
    Explanations

    references to power and performance in a competitive context

    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
    readcr
    -0.18
     getpid
    -0.17
    getManager
    -0.14
    landa
    -0.14
    igham
    -0.14
     Wich
    -0.14
    smarty
    -0.14
    .DEFINE
    -0.14
    edException
    -0.14
    niž
    -0.14
    POSITIVE LOGITS
     power
    0.33
    power
    0.27
     Power
    0.25
     POWER
    0.25
    Power
    0.24
    -power
    0.24
    _power
    0.23
     slug
    0.22
    (power
    0.21
    POWER
    0.21
    Activations Density 0.013%

    No Known Activations