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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    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. 23586
    Prev
    Next
    INDEX
    Explanations

    intensifiers or modifiers that express degree or extent

    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
    qus
    -0.16
    imore
    -0.16
    ugo
    -0.15
     ApplicationException
    -0.15
    uars
    -0.15
    ìĨ
    -0.14
    maf
    -0.13
    AREST
    -0.13
     Schwe
    -0.13
    ãĤ·ãĤ¢
    -0.13
    POSITIVE LOGITS
     much
    0.60
    much
    0.44
     Much
    0.42
     MUCH
    0.41
    Much
    0.40
     mucho
    0.33
     many
    0.30
     veel
    0.24
     viel
    0.24
     beaucoup
    0.24
    Activations Density 0.036%

    No Known Activations