© 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. 7521
    Prev
    Next
    INDEX
    Explanations

    instances where something is being described or pointed out in a specific location (e.g., "shown in the image below")

    oai_token-act-pair · gpt-3.5-turbo

    references to visual elements or images in the content

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    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
     metic
    -0.65
    orous
    -0.63
    emale
    -0.63
    exempt
    -0.63
    cision
    -0.63
    ococ
    -0.58
     Elim
    -0.58
     Aust
    -0.56
    icist
    -0.56
    ttle
    -0.56
    POSITIVE LOGITS
     below
    1.44
     above
    1.36
    below
    1.32
     screenshot
    1.25
    above
    1.25
     slideshow
    1.21
     gallery
    1.15
     links
    1.14
     screenshots
    1.12
     sidebar
    1.09
    Activations Density 0.198%

    No Known Activations