© 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. 1-RES-JB
    11. 11609
    Prev
    Next
    INDEX
    Explanations

    dates specified as days of the week

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of the day "Monday."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 1-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    1
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ript
    -0.96
    ebook
    -0.90
    unning
    -0.86
    ifled
    -0.84
    NRS
    -0.82
    arist
    -0.80
    keye
    -0.80
    CHQ
    -0.78
    abet
    -0.77
    requ
    -0.73
    POSITIVE LOGITS
     morning
    1.48
     mornings
    1.40
     afternoon
    1.37
     night
    1.26
     evening
    1.26
    morning
    1.18
     Night
    1.15
     nights
    1.14
     evenings
    1.07
     Monday
    1.04
    Activations Density 0.019%

    No Known Activations