© 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. 25528
    Prev
    Next
    INDEX
    Explanations

    phrases that express uncertainty or conditional suggestions

    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
    'gc
    -0.18
    mÃŃt
    -0.17
    .scalablytyped
    -0.15
    utto
    -0.15
    iesz
    -0.15
    ancy
    -0.14
    estre
    -0.14
     DropIndex
    -0.14
    bach
    -0.14
    iquer
    -0.14
    POSITIVE LOGITS
     simply
    0.63
     Simply
    0.54
    Simply
    0.48
     simplement
    0.45
     just
    0.44
     пÑĢоÑģÑĤо
    0.41
     einfach
    0.38
    just
    0.35
     simple
    0.34
     simples
    0.33
    Activations Density 0.027%

    No Known Activations