© 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. 0-RES-JB
    11. 4780
    Prev
    Next
    INDEX
    Explanations

    words signaling a contrast or change in direction

    oai_token-act-pair · gpt-3.5-turbo

    the repeated phrasing "while" to denote contrast or ongoing conditions

    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
     instinct
    -0.65
     defin
    -0.64
     gravy
    -0.63
     fine
    -0.62
     answer
    -0.62
     correct
    -0.61
     optim
    -0.61
     ideal
    -0.60
     general
    -0.59
     kind
    -0.59
    POSITIVE LOGITS
     While
    2.76
     Whilst
    2.42
     Though
    2.11
    While
    2.03
     Although
    2.01
     Despite
    1.89
     Whereas
    1.79
     During
    1.78
     Meanwhile
    1.74
    while
    1.70
    Activations Density 0.040%

    No Known Activations