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

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 21608
    Prev
    Next
    INDEX
    Explanations

    terms and concepts related to metrics, support systems, and indicators of performance or social constructs

    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
    à¹īà¹Ģà¸Ľ
    -0.15
     als
    -0.15
     ako
    -0.14
    berger
    -0.14
     such
    -0.14
    ham
    -0.14
    -ÑĤаки
    -0.14
     SUCH
    -0.13
    ifr
    -0.13
    è¿Ļæł·çļĦ
    -0.13
    POSITIVE LOGITS
     for
    0.32
     длÑı
    0.25
     bagi
    0.23
    	for
    0.22
    for
    0.21
    สำหร
    0.20
     dla
    0.19
     für
    0.17
     capable
    0.17
    длÑı
    0.17
    Activations Density 0.195%

    No Known Activations