© 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. 33356
    Prev
    Next
    INDEX
    Explanations

    discussions about taxes, especially marginal tax rates and surcharges on high-income earners

    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
    سÛĮ
    -0.13
    ìŀ¬
    -0.13
    vard
    -0.13
    ewe
    -0.13
    elah
    -0.13
     Hed
    -0.13
     gezocht
    -0.13
    iggins
    -0.12
    [sizeof
    -0.12
    edom
    -0.12
    POSITIVE LOGITS
     applies
    0.26
     effective
    0.22
     apply
    0.22
     applicable
    0.21
     applying
    0.20
    éģ©çĶ¨
    0.20
     Applies
    0.20
     applic
    0.19
     ìłģìļ©
    0.19
     cap
    0.19
    Activations Density 0.077%

    No Known Activations