© 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. 40693
    Prev
    Next
    INDEX
    Explanations

    references to economic conditions and their impacts on employment and housing

    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
    iah
    -0.16
    ahan
    -0.15
    å°ĸ
    -0.15
     оз
    -0.15
    ouch
    -0.14
    ุà¸Ĺà¸ĺ
    -0.14
    apus
    -0.14
    å¦ĥ
    -0.14
    /ay
    -0.14
    .Persistent
    -0.14
    POSITIVE LOGITS
     recovery
    0.29
     strong
    0.27
     Recovery
    0.27
     recovering
    0.26
     recover
    0.26
     recovered
    0.26
     Recover
    0.25
     robust
    0.24
     improving
    0.23
     Rob
    0.22
    Activations Density 0.046%

    No Known Activations