© 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. 55334
    Prev
    Next
    INDEX
    Explanations

    sentences that contain strong emotional or dramatic punctuation

    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
    serrat
    -0.14
    olla
    -0.14
    .tbl
    -0.14
    ullo
    -0.14
    миÑĢ
    -0.14
    hani
    -0.13
    asurer
    -0.13
    hb
    -0.13
    ancode
    -0.13
    ÏįÏĦε
    -0.13
    POSITIVE LOGITS
     except
    0.54
     Except
    0.51
    Except
    0.50
    except
    0.47
     problem
    0.44
     Problem
    0.40
     Trouble
    0.39
     Unfortunately
    0.38
     trouble
    0.37
    Unfortunately
    0.37
    Activations Density 0.059%

    No Known Activations