© 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
    2. Julius Han · Gated SAE for Llama3-8B-Instruct
    3. Llama3-8B-IT
    4. Residual Stream
    5. 25-RES-JH
    6. 30083
    Prev
    Next
    INDEX
    Explanations

    present participles and action-related verbs in various contexts

    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
    NotificationCenter
    -0.17
    ุà¸ģ
    -0.16
     twice
    -0.15
    ัà¸Ļà¹Ħà¸Ķ
    -0.15
     entire
    -0.15
    ãģĻãģ¹ãģ¦
    -0.15
     caÅĤ
    -0.15
    ude
    -0.15
    upon
    -0.14
    eras
    -0.14
    POSITIVE LOGITS
    /testing
    0.23
    /loading
    0.23
    /logging
    0.21
     away
    0.19
    ly
    0.17
     another
    0.16
     оÑĩеÑĢед
    0.16
    reating
    0.16
    ÃŃcÃŃ
    0.16
    951
    0.16
    Activations Density 0.081%

    No Known Activations