© 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 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. 42598
    Prev
    Next
    INDEX
    Explanations

    references to business operations and continuity

    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
     finally
    -0.20
     Finally
    -0.17
    ç»Īäºİ
    -0.16
    finally
    -0.16
    505
    -0.14
    vely
    -0.14
    _rq
    -0.14
     Fried
    -0.14
    Finally
    -0.14
     elusive
    -0.14
    POSITIVE LOGITS
     continue
    0.34
     continues
    0.31
     continuing
    0.30
     continu
    0.28
    continue
    0.28
     continued
    0.28
     Continue
    0.28
     continuation
    0.28
     retained
    0.26
    Continue
    0.26
    Activations Density 0.037%

    No Known Activations