© 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. 25079
    Prev
    Next
    INDEX
    Explanations

    terms related to coding and scoring in research studies

    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
    iano
    -0.15
    olas
    -0.14
    ande
    -0.14
     Bund
    -0.14
     instincts
    -0.13
     supplement
    -0.13
    ome
    -0.13
    airs
    -0.13
    ak
    -0.13
     Ak
    -0.13
    POSITIVE LOGITS
     blind
    0.42
    -blind
    0.40
    blind
    0.38
     Blind
    0.35
     blindness
    0.31
     blinded
    0.31
    -mask
    0.24
     bl
    0.24
     blindly
    0.23
     double
    0.22
    Activations Density 0.025%

    No Known Activations