Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 0-RES-JB
    6. 956
    Prev
    Next
    INDEX
    Explanations

    words related to windows, such as "windows", "illuminated windows", "digital windows", and "household projects like windows"

    oai_token-act-pair · gpt-3.5-turbo

    references to windows

    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
    AI
    -0.71
     Tribe
    -0.69
     Singer
    -0.69
    iety
    -0.67
    ylum
    -0.66
    ayer
    -0.65
    ieth
    -0.63
    uren
    -0.62
    OTAL
    -0.62
    Ä«
    -0.62
    POSITIVE LOGITS
     windows
    3.90
     window
    2.64
    windows
    2.36
     Window
    2.07
    window
    1.89
     doors
    1.74
    Window
    1.62
     windshield
    1.58
     roofs
    1.46
     ceilings
    1.45
    Activations Density 0.014%

    No Known Activations