© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 42437
    Prev
    Next
    INDEX
    Explanations

    references to sports and physical performance, particularly focusing on player statistics and impacts

    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
     Fault
    -0.17
    atk
    -0.16
     flaw
    -0.16
     fault
    -0.15
    ppo
    -0.15
    Fault
    -0.14
     crippling
    -0.14
     geil
    -0.14
    ecess
    -0.14
    lops
    -0.14
    POSITIVE LOGITS
     headaches
    0.38
     headache
    0.35
     grief
    0.35
     trouble
    0.29
     problems
    0.29
     Problems
    0.25
     heart
    0.25
     hassle
    0.24
    Trou
    0.23
     nightmares
    0.23
    Activations Density 0.063%

    No Known Activations