Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Julius Han · Gated SAE for Llama3-8B-Instruct
    3. Llama3-8B-IT
    4. Residual Stream
    5. 25-RES-JH
    6. 63039
    Prev
    Next
    INDEX
    Explanations

    references to nautical or transportation-related items

    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
    opping
    -0.17
    erea
    -0.17
    ailable
    -0.15
    utow
    -0.15
    à¹Ĥà¸Ķ
    -0.14
    対
    -0.14
    alley
    -0.14
     polož
    -0.13
    achen
    -0.13
    iling
    -0.13
    POSITIVE LOGITS
     containing
    0.62
     full
    0.45
     contain
    0.42
     carrying
    0.41
     filled
    0.41
     contains
    0.41
    Containing
    0.38
    full
    0.37
    -full
    0.37
     chứa
    0.36
    Activations Density 0.035%

    No Known Activations