© 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. 11-RES-JB
    11. 7544
    Prev
    Next
    INDEX
    Explanations

    holiday greetings

    oai_token-act-pair · gpt-3.5-turbo

    sentences ending with punctuation

    oai_token-act-pair · gpt-4o-miniTriggered by @bot

    Period at the end of sentences about big changes, especially sentences related to climate.

    Explanation Uploaded by User
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     challeng
    -0.83
     dimensional
    -0.80
     multif
    -0.78
     advoc
    -0.78
     satell
    -0.77
     predec
    -0.77
     immersion
    -0.73
     glim
    -0.73
     structured
    -0.72
     peripher
    -0.71
    POSITIVE LOGITS
     Photo
    1.30
     Photograph
    1.26
     Shutterstock
    1.20
     Unless
    1.10
     Especially
    1.09
     PHOTO
    1.09
     Instead
    1.09
     But
    1.07
     Image
    1.07
     |
    1.06
    Activations Density 0.518%

    No Known Activations