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. 23440
    Prev
    Next
    INDEX
    Explanations

    marks of formatting or structure within the text

    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
    .dtd
    -0.14
    ).__
    -0.13
    embre
    -0.13
    Descricao
    -0.13
    oded
    -0.13
    æķı
    -0.12
    باØŃ
    -0.12
    наÑĩала
    -0.12
    inge
    -0.12
    ìļ´ëį°
    -0.12
    POSITIVE LOGITS
     References
    0.35
    References
    0.31
     references
    0.30
    references
    0.25
     reference
    0.22
     REFERENCES
    0.21
    åıĤèĢĥ
    0.19
     Notes
    0.19
     External
    0.19
     åıĤèĢĥ
    0.19
    Activations Density 0.086%

    No Known Activations