© 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. 30094
    Prev
    Next
    INDEX
    Explanations

    instances of the word "as."

    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
    oretical
    -0.16
    isÃŃ
    -0.15
    ories
    -0.15
    gether
    -0.15
    Ñħод
    -0.15
    evity
    -0.15
    odore
    -0.15
    eous
    -0.15
    orem
    -0.15
    елем
    -0.15
    POSITIVE LOGITS
    cert
    0.30
    paragus
    0.29
    phy
    0.27
    sembl
    0.26
    yl
    0.25
    cribed
    0.24
     opposed
    0.24
    cribing
    0.23
    sembles
    0.22
    ides
    0.22
    Activations Density 0.235%

    No Known Activations