© 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. 14043
    Prev
    Next
    INDEX
    Explanations

    references to early stages or historical periods

    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
    older
    -0.20
     older
    -0.17
    amp
    -0.16
    ãģĿãģ®ä»ĸ
    -0.15
    akan
    -0.15
     latest
    -0.14
     ultimate
    -0.14
     ancient
    -0.14
    odu
    -0.14
    ching
    -0.14
    POSITIVE LOGITS
    -stage
    0.30
     stages
    0.27
    -warning
    0.25
     days
    0.24
    days
    0.21
    -middle
    0.20
    -days
    0.20
    -care
    0.20
     stage
    0.19
     еÑĤап
    0.17
    Activations Density 0.039%

    No Known Activations