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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 29756
    Prev
    Next
    INDEX
    Explanations

    expressions of desire for change or new experiences

    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
     Many
    -0.19
     many
    -0.19
     often
    -0.17
    HorizontalAlignment
    -0.16
    many
    -0.15
     countless
    -0.15
     quite
    -0.15
     almost
    -0.15
     nearly
    -0.15
    Many
    -0.14
    POSITIVE LOGITS
     something
    0.31
     somehow
    0.29
    something
    0.28
    Something
    0.24
    omething
    0.24
     möglich
    0.24
     Something
    0.21
     preferably
    0.21
     someday
    0.20
     quelque
    0.20
    Activations Density 0.053%

    No Known Activations