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. 23677
    Prev
    Next
    INDEX
    Explanations

    concepts related to testing and detection

    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
    AINS
    -0.17
    оÑĢон
    -0.15
    _intent
    -0.15
     engineering
    -0.14
     Conduct
    -0.14
    ewn
    -0.14
    ót
    -0.14
     Engineering
    -0.14
    ctrine
    -0.13
    ¨¡
    -0.13
    POSITIVE LOGITS
     pick
    0.45
     picking
    0.45
     picked
    0.41
     picks
    0.40
    pick
    0.40
     pickup
    0.38
     Pick
    0.38
    Pick
    0.37
     PICK
    0.35
     catch
    0.33
    Activations Density 0.027%

    No Known Activations