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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 63550
    Prev
    Next
    INDEX
    Explanations

    references to structural elements and relationships within various contexts, particularly in narrative or observational frameworks

    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
    oref
    -0.14
    ead
    -0.13
     Bentley
    -0.13
    ansom
    -0.13
    icont
    -0.13
    jak
    -0.13
    hin
    -0.13
    rud
    -0.13
    yp
    -0.13
     ki
    -0.12
    POSITIVE LOGITS
     itself
    0.50
     being
    0.47
     concerned
    0.41
    being
    0.40
     Being
    0.39
    Being
    0.39
     involved
    0.35
    被
    0.30
     sendo
    0.30
     they
    0.30
    Activations Density 0.266%

    No Known Activations