© 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. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 10-RES-JB
    11. 2972
    Prev
    Next
    INDEX
    Explanations

    words related to specific terms or concepts, such as 'God', 'marriage', 'fake news', 'civil', 'assassinate', and 'socialist'

    oai_token-act-pair · gpt-3.5-turbo

    terms that have significant cultural, social, or political implications

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    10
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     livest
    -0.62
     originals
    -0.61
     DEC
    -0.60
     traject
    -0.59
     incentives
    -0.59
     nightly
    -0.59
     archives
    -0.58
     DVDs
    -0.58
     Paige
    -0.58
     Exhibit
    -0.58
    POSITIVE LOGITS
    âĢİ
    1.05
     refers
    0.97
     denotes
    0.94
    É
    0.92
    ت
    0.89
    د
    0.88
    ر
    0.86
     implies
    0.85
    ب
    0.84
    notation
    0.83
    Activations Density 0.200%

    No Known Activations