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. 23543
    Prev
    Next
    INDEX
    Explanations

    expressions related to urgency and quick action

    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
     finally
    -0.27
     slowly
    -0.23
     eventually
    -0.22
     Finally
    -0.20
    Finally
    -0.20
     gradually
    -0.20
    finally
    -0.19
    slow
    -0.19
     Slow
    -0.19
     slow
    -0.18
    POSITIVE LOGITS
    aneously
    0.26
    urt
    0.20
     chóng
    0.20
    aneous
    0.20
     (<
    0.18
    est
    0.17
     once
    0.17
    ane
    0.16
    amente
    0.16
    otropic
    0.15
    Activations Density 0.039%

    No Known Activations