© 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
    3. Julius Han · Gated SAE for Llama3-8B-Instruct
    5. Llama3-8B-IT
    7. Residual Stream
    9. 25-RES-JH
    11. 9172
    Prev
    Next
    INDEX
    Explanations

    punctuation and sentence structure

    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
    estruction
    -0.13
    dev
    -0.13
    (Source
    -0.13
    oundation
    -0.13
    osci
    -0.13
    aeda
    -0.13
    ĶĦ
    -0.12
    Reducers
    -0.12
    umd
    -0.12
    .bat
    -0.12
    POSITIVE LOGITS
     I
    0.19
     maybe
    0.18
     It
    0.18
     He
    0.17
     then
    0.16
     Inst
    0.16
     Then
    0.16
    Inst
    0.15
     There
    0.15
     His
    0.15
    Activations Density 0.278%

    No Known Activations