© 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. 55060
    Prev
    Next
    INDEX
    Explanations

    instances of the letter 'b' and contractions involving 'm' and 'not.'

    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
    ãĥ¼ãĥĩ
    -0.15
     indeb
    -0.14
    à¹īำ
    -0.14
    uet
    -0.14
    enet
    -0.14
     itself
    -0.14
    丸
    -0.13
    uki
    -0.13
    uzu
    -0.13
    abis
    -0.13
    POSITIVE LOGITS
    woke
    0.17
    pired
    0.16
     поба
    0.15
    ãģ¾ãģŁ
    0.15
    ANNOT
    0.15
    vanished
    0.15
    unfold
    0.14
    eded
    0.14
    หาย
    0.14
     ourselves
    0.14
    Activations Density 0.001%

    No Known Activations