© 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
    2. Julius Han · Gated SAE for Llama3-8B-Instruct
    3. Llama3-8B-IT
    4. Residual Stream
    5. 25-RES-JH
    6. 54935
    Prev
    Next
    INDEX
    Explanations

    references to what individuals have stated or claimed

    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
    bai
    -0.16
    anton
    -0.15
    otton
    -0.15
    íĴĪ
    -0.15
    ovah
    -0.14
    лÑİб
    -0.14
    _UNIX
    -0.14
    лад
    -0.13
    ÑĢана
    -0.13
    /AP
    -0.13
    POSITIVE LOGITS
     nothing
    0.30
     say
    0.27
     something
    0.27
     words
    0.27
    nothing
    0.26
     NOTHING
    0.25
     thing
    0.24
     things
    0.24
     Nothing
    0.23
    Nothing
    0.22
    Activations Density 0.021%

    No Known Activations