© 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. 9449
    Prev
    Next
    INDEX
    Explanations

    instances of the word "Paris" with high activation values

    oai_token-act-pair · gpt-3.5-turbo

    mentions of the city of Paris

    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
     reliever
    -0.88
    :(
    -0.76
    ullah
    -0.70
     indic
    -0.69
     bob
    -0.67
     bullpen
    -0.66
    UD
    -0.66
    Shield
    -0.64
     fielded
    -0.63
     hitters
    -0.63
    POSITIVE LOGITS
     Paris
    3.68
    Paris
    3.30
     Copenhagen
    2.02
     France
    1.90
    France
    1.79
     Brussels
    1.77
     Marse
    1.76
     Cologne
    1.67
     Vienna
    1.66
     Lisbon
    1.64
    Activations Density 0.015%

    No Known Activations