© 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. 4-RES-JB
    11. 20257
    Prev
    Next
    INDEX
    Explanations

    numerical references to the ordinal position 'first' within a context

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "first" followed by a numerical value indicating a time period or event

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

    No Comments

    Negative Logits
    tics
    -0.77
    ingen
    -0.71
    bos
    -0.68
    mbuds
    -0.67
     Canaver
    -0.64
    Sport
    -0.62
    mund
    -0.62
     Motorsport
    -0.61
    leaders
    -0.61
    sav
    -0.60
    POSITIVE LOGITS
     responders
    1.03
     baseman
    0.99
     installment
    0.95
     lady
    0.92
     blush
    0.89
     batch
    0.86
     iteration
    0.83
     incarnation
    0.81
     attempt
    0.81
     ascent
    0.81
    Activations Density 0.085%

    No Known Activations