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

    Neuronpedia

    Jacobian LensNEW
    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. 2-RES-JB
    11. 22456
    Prev
    Next
    INDEX
    Explanations

    words related to isotopes, nuclear fission, and radiation

    oai_token-act-pair · gpt-3.5-turbo

    references to isotopes and nuclear science

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 2-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    2
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    ????????
    -0.83
     Lawyers
    -0.81
     Quadro
    -0.78
    Rogue
    -0.78
    à
    -0.76
    Sexual
    -0.72
     NCT
    -0.70
    advertisement
    -0.69
    Snow
    -0.68
    DT
    -0.68
    POSITIVE LOGITS
     isot
    1.19
    obyl
    0.87
    illery
    0.85
     decay
    0.85
     petroleum
    0.82
     physicists
    0.81
     plutonium
    0.81
    roleum
    0.80
    atsu
    0.79
     hydrogen
    0.79
    Activations Density 0.018%

    No Known Activations