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

    Neuronpedia

    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. 15411
    Prev
    Next
    INDEX
    Explanations

    key phrases or concepts related to anomalies and database statuses

    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
     really
    -0.26
     like
    -0.21
     big
    -0.20
     lots
    -0.19
     just
    -0.19
     partly
    -0.18
     pretty
    -0.18
     bigger
    -0.17
     get
    -0.17
     even
    -0.17
    POSITIVE LOGITS
     attempting
    0.18
    Attempting
    0.17
    fuck
    0.17
    Fuck
    0.16
     fucks
    0.16
    attempt
    0.16
     FUCK
    0.15
    elope
    0.15
     Currently
    0.15
     âĸĪ
    0.15
    Activations Density 0.003%

    No Known Activations