© 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. 63386
    Prev
    Next
    INDEX
    Explanations

    expressions of disappointment and the organization of events

    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
    оÑĩÑĮ
    -0.15
     discrim
    -0.15
     suff
    -0.15
     soon
    -0.15
    Ä°T
    -0.14
    à¹Ĥย
    -0.14
     Soon
    -0.13
    анÑĤ
    -0.13
    رش
    -0.13
    sink
    -0.13
    POSITIVE LOGITS
     staying
    0.22
     stayed
    0.21
     stay
    0.21
     drive
    0.20
     arrive
    0.20
     arriv
    0.20
     arriving
    0.20
     arrived
    0.20
     driving
    0.20
     drove
    0.19
    Activations Density 0.177%

    No Known Activations