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. 23722
    Prev
    Next
    INDEX
    Explanations

    instances of the word "compare" and its variations

    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
    çĦ¶
    -0.19
    ereotype
    -0.17
    anna
    -0.16
    amping
    -0.15
    .nz
    -0.15
    еÑİ
    -0.14
    ough
    -0.14
    elt
    -0.14
    ivia
    -0.14
    imore
    -0.14
    POSITIVE LOGITS
     apples
    0.27
     favor
    0.22
     unfavor
    0.22
    favor
    0.22
     favour
    0.20
     between
    0.19
    rios
    0.18
    isons
    0.18
     notes
    0.18
    ãģ¹
    0.17
    Activations Density 0.023%

    No Known Activations