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. 23393
    Prev
    Next
    INDEX
    Explanations

    phrases that reference a variety or range of items or concepts

    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
     portions
    -0.18
     certain
    -0.17
     or
    -0.17
     Certain
    -0.17
    lant
    -0.17
     series
    -0.16
     portion
    -0.16
     bestimm
    -0.14
    Certain
    -0.14
     Portions
    -0.14
    POSITIVE LOGITS
     possibilities
    0.23
     different
    0.21
     possible
    0.20
     possibile
    0.19
    different
    0.19
     topics
    0.18
     styles
    0.18
     POSS
    0.17
    åı¯èĥ½æĢ§
    0.17
     colors
    0.16
    Activations Density 0.098%

    No Known Activations