© 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 EvalsExportsAPI 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. 42465
    Prev
    Next
    INDEX
    Explanations

    customer reviews expressing satisfaction or dissatisfaction with products

    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
    ần
    -0.16
     Commit
    -0.15
    .yy
    -0.15
    andra
    -0.15
    itas
    -0.15
    ãĥ³ãĥķ
    -0.15
    afone
    -0.14
    ÙĦÙħاÙĨ
    -0.14
    коз
    -0.14
     collar
    -0.14
    POSITIVE LOGITS
     notice
    0.20
     overall
    0.18
    overall
    0.17
     notices
    0.17
     purchased
    0.17
     dock
    0.17
     paid
    0.17
    notice
    0.16
    Overall
    0.16
     experienced
    0.16
    Activations Density 0.098%

    No Known Activations