© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 0-RES-JB
    11. 1295
    Prev
    Next
    INDEX
    Explanations

    references to women, specifically the term "ladies."

    oai_token-act-pair · gpt-3.5-turbo

    references to 'ladies' and 'gentlemen.'

    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
    awi
    -0.78
    asca
    -0.69
    inez
    -0.68
    obo
    -0.67
     fragmented
    -0.65
    ingu
    -0.65
    awed
    -0.65
    aq
    -0.64
    TPS
    -0.62
     Flood
    -0.62
    POSITIVE LOGITS
     ladies
    3.45
     Ladies
    2.80
     gentlemen
    2.79
     gentleman
    2.49
     lady
    1.96
     Gentle
    1.95
     Gentleman
    1.92
    adies
    1.67
    entle
    1.62
     sir
    1.38
    Activations Density 0.020%

    No Known Activations