© 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. 8-RES-JB
    11. 7882
    Prev
    Next
    INDEX
    Explanations

    phrases related to information about specific topics or entities, possibly indicated by the word "About" followed by keywords

    oai_token-act-pair · gpt-3.5-turbo

    phrases indicating the topic or subject of a discussion

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    BT
    -0.71
     jaws
    -0.61
    ãģ®ç
    -0.61
     slightest
    -0.61
    cious
    -0.60
    NEY
    -0.59
    OTUS
    -0.57
    rift
    -0.57
     sewer
    -0.56
     tending
    -0.56
    POSITIVE LOGITS
     Us
    0.95
     Languages
    0.78
     Seller
    0.74
     Paste
    0.73
    iframe
    0.73
     Comments
    0.68
     Develop
    0.66
     Definitions
    0.66
     References
    0.64
    aleb
    0.63
    Activations Density 0.029%

    No Known Activations