© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    2. Julius Han · Gated SAE for Llama3-8B-Instruct
    3. Llama3-8B-IT
    4. Residual Stream
    5. 25-RES-JH
    6. 29064
    Prev
    Next
    INDEX
    Explanations

    phrases related to financial statistics and accounting percentages

    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
    undred
    -0.15
     Few
    -0.14
     Hundred
    -0.13
    ildo
    -0.13
    eldom
    -0.13
    Ùĩار
    -0.13
    andard
    -0.12
    .nr
    -0.12
    initely
    -0.12
    uru
    -0.12
    POSITIVE LOGITS
     half
    0.25
     about
    0.24
     largest
    0.24
     lion
    0.23
     majority
    0.23
     around
    0.22
     roughly
    0.20
     lions
    0.20
     HALF
    0.19
     nearly
    0.19
    Activations Density 0.100%

    No Known Activations