© 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. 41889
    Prev
    Next
    INDEX
    Explanations

    phrases related to the concept of being "too big to fail."

    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
    éħ
    -0.17
     Liberation
    -0.15
    éĴŁ
    -0.15
    esh
    -0.14
    raphics
    -0.14
    ç´¹
    -0.14
     Ramirez
    -0.14
     cellFor
    -0.14
     Judge
    -0.14
     ResourceType
    -0.13
    POSITIVE LOGITS
     touch
    0.19
     touched
    0.17
    ailability
    0.17
     Touch
    0.16
    ippi
    0.15
    quine
    0.15
    oplast
    0.15
    -touch
    0.15
     touches
    0.15
    touch
    0.15
    Activations Density 0.013%

    No Known Activations