© 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. 29550
    Prev
    Next
    INDEX
    Explanations

    references to contrasting or binary concepts

    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
    uhl
    -0.15
    combination
    -0.15
    .native
    -0.14
    жÑĥ
    -0.14
    aze
    -0.14
     combination
    -0.14
     intersections
    -0.14
     Various
    -0.14
    ãĥĬãĥ«
    -0.14
     massaggi
    -0.14
    POSITIVE LOGITS
     halves
    0.35
     sides
    0.34
     worlds
    0.28
     camps
    0.25
     opposing
    0.24
     sets
    0.24
     parts
    0.24
     extremes
    0.23
     Worlds
    0.22
     wings
    0.21
    Activations Density 0.106%

    No Known Activations