© 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. 3-RES-JB
    11. 5828
    Prev
    Next
    INDEX
    Explanations

    references to different levels in a building, particularly related to movement

    oai_token-act-pair · gpt-3.5-turbo

    references to locations, particularly levels of a building such as upstairs and downstairs

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    aceous
    -0.71
    tex
    -0.69
    ck
    -0.68
    nesota
    -0.66
    itar
    -0.65
    eming
    -0.65
    elson
    -0.64
    cks
    -0.63
    cking
    -0.62
    itarian
    -0.61
    POSITIVE LOGITS
    stairs
    1.59
     upstairs
    1.17
     downstairs
    1.16
     stairs
    0.96
    bledon
    0.87
     adolesc
    0.80
     challeng
    0.79
     bedrooms
    0.79
     proport
    0.75
     Abbey
    0.75
    Activations Density 0.011%

    No Known Activations