© 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. 0-RES-JB
    11. 10434
    Prev
    Next
    INDEX
    Explanations

    specific patterns related to programming languages, likely focusing on the syntax of C++ with particular attention to identifiers and programming constructs

    oai_token-act-pair · gpt-3.5-turbo

    technical jargon or specific codes related to programming

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     trunk
    -0.68
     anat
    -0.67
     observers
    -0.64
     population
    -0.63
     orderly
    -0.63
     hem
    -0.61
     observer
    -0.60
     holder
    -0.60
     weighing
    -0.59
     altern
    -0.59
    POSITIVE LOGITS
    xx
    3.91
    xxx
    2.13
    xxxx
    1.54
    xi
    1.36
    xy
    1.30
    xus
    1.21
    lez
    1.14
    eger
    1.07
    x
    1.04
     xx
    1.00
    Activations Density 0.017%

    No Known Activations