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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 25450
    Prev
    Next
    INDEX
    Explanations

    significant references to individuals or authority figures

    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
    ãĥ¬ãĤ¹
    -0.14
    //{{
    -0.13
     recru
    -0.13
    alsex
    -0.12
    BuilderInterface
    -0.12
    ýt
    -0.12
    ervices
    -0.12
     BaseService
    -0.12
     unimagin
    -0.12
    áp
    -0.12
    POSITIVE LOGITS
     same
    0.20
     whole
    0.19
    odor
    0.18
     difference
    0.18
     cumulative
    0.18
     effect
    0.18
     more
    0.18
     entire
    0.17
     word
    0.17
     latter
    0.17
    Activations Density 0.411%

    No Known Activations