Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 3-RES-JB
    6. 18889
    Prev
    Next
    INDEX
    Explanations

    lines of text starting and ending with specific characters ('Ċ' and '|endoftext|') and containing various criteria or conditions

    oai_token-act-pair · gpt-3.5-turbo

    list-like structures or enumerations in text

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.hook_resid_pre
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    hement
    -0.83
    senal
    -0.79
    ataka
    -0.71
     Thornton
    -0.68
     emancipation
    -0.67
    ĸļ
    -0.66
     lifes
    -0.65
    aido
    -0.63
     Guth
    -0.61
    orate
    -0.61
    POSITIVE LOGITS
    Interstitial
    0.90
    ;;
    0.85
    Liter
    0.82
    Introdu
    0.79
    FIN
    0.77
    âĸł
    0.76
    Height
    0.76
    Ur
    0.75
    Ingredients
    0.75
    âľ
    0.74
    Activations Density 0.230%

    No Known Activations