© 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. 10090
    Prev
    Next
    INDEX
    Explanations

    the name "Curtis"

    oai_token-act-pair · gpt-3.5-turbo

    the names of individuals, particularly "Curtis" and references to "Sag" and "Vega"

    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
     Scrolls
    -0.88
    Ùİ
    -0.79
    quished
    -0.78
     Wear
    -0.76
     fri
    -0.75
    Ĥª
    -0.75
     Mour
    -0.73
     Flavoring
    -0.73
     rim
    -0.72
     Dw
    -0.71
    POSITIVE LOGITS
     Curtis
    1.86
     Vega
    1.59
     Sag
    1.55
     Buster
    1.38
     Pascal
    1.37
    rax
    1.35
    ilot
    1.23
    ravis
    1.19
     Miranda
    1.18
     Kass
    1.10
    Activations Density 0.054%

    No Known Activations