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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsBlog/PodcastSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlog/PodcastGitHubSlackTwitterContact
    1. Home
    2. Julius Han · Gated SAE for Llama3-8B-Instruct
    3. Llama3-8B-IT
    4. Residual Stream
    5. 25-RES-JH
    6. 23602
    Prev
    Next
    INDEX
    Explanations

    file extensions and technical terms related to file types and data formats

    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
    yonel
    -0.19
    ettes
    -0.18
    dz
    -0.18
    ovich
    -0.17
    off
    -0.17
    etting
    -0.17
    earch
    -0.17
    ect
    -0.17
    eme
    -0.16
    tingham
    -0.16
    POSITIVE LOGITS
    s
    0.34
    l
    0.26
    ows
    0.24
    ustr
    0.22
    kommen
    0.21
    owing
    0.20
    llll
    0.20
    owy
    0.19
    (strict
    0.19
    iams
    0.18
    Activations Density 0.019%

    No Known Activations