© 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. 29624
    Prev
    Next
    INDEX
    Explanations

    phrases indicating reduction or diminishing amounts

    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
    adera
    -0.15
    opic
    -0.15
    isses
    -0.14
    eh
    -0.14
    acher
    -0.14
     Bounty
    -0.14
    oug
    -0.14
    ady
    -0.14
    ouser
    -0.14
    ha
    -0.14
    POSITIVE LOGITS
    ening
    0.34
    ened
    0.34
    -than
    0.25
     likely
    0.24
    ens
    0.21
     inclined
    0.20
    ons
    0.20
    sons
    0.18
    ONS
    0.18
     Likely
    0.17
    Activations Density 0.034%

    No Known Activations