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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 30203
    Prev
    Next
    INDEX
    Explanations

    references to mathematical concepts or visualizations

    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
     Masc
    -0.17
     Lump
    -0.15
    orizontal
    -0.14
    _chunks
    -0.14
    Scoped
    -0.14
     hạng
    -0.14
     Crud
    -0.13
     deb
    -0.13
     dummy
    -0.13
     Dickens
    -0.13
    POSITIVE LOGITS
     patterns
    0.30
     fract
    0.28
     Patterns
    0.27
     pattern
    0.26
    Patterns
    0.25
     Mand
    0.25
     tess
    0.25
    patterns
    0.24
     Pattern
    0.24
    pattern
    0.23
    Activations Density 0.140%

    No Known Activations