© 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. 41108
    Prev
    Next
    INDEX
    Explanations

    actions and states related to changes and adjustments

    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
    unger
    -0.14
    itis
    -0.14
    ãģłãģij
    -0.14
     because
    -0.14
    Ùī
    -0.13
    illions
    -0.13
     too
    -0.13
    ãģ§ãģįãĤĭ
    -0.13
    cai
    -0.13
     deline
    -0.13
    POSITIVE LOGITS
     pretty
    0.28
     quite
    0.26
    pretty
    0.25
     somewhat
    0.24
     fairly
    0.24
     khá
    0.20
    quite
    0.20
     rather
    0.20
    rather
    0.19
     Quite
    0.18
    Activations Density 0.282%

    No Known Activations