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. 23147
    Prev
    Next
    INDEX
    Explanations

    phrases indicating time or temporal context in political discussions

    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
    LIKELY
    -0.16
    æ¦
    -0.15
     Wid
    -0.15
     rib
    -0.15
     Spacer
    -0.14
    .persistent
    -0.14
    оÑĤÑĢеб
    -0.14
    µ
    -0.14
    iske
    -0.14
     Fraction
    -0.14
    POSITIVE LOGITS
     cycle
    0.45
     Cycle
    0.38
     cycles
    0.38
    -cycle
    0.36
    cycle
    0.34
    Cycle
    0.34
    cycles
    0.29
    _cycle
    0.27
    åij¨æľŁ
    0.27
    .cycle
    0.26
    Activations Density 0.037%

    No Known Activations