© 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. 0-RES-JB
    11. 9624
    Prev
    Next
    INDEX
    Explanations

    words related to cleanliness or lack thereof

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "dirty" in various contexts

    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
    lement
    -0.75
    conom
    -0.71
    OPLE
    -0.68
     Publishers
    -0.66
     Kaplan
    -0.66
    clus
    -0.66
    cephal
    -0.66
     remission
    -0.64
    haps
    -0.63
    roups
    -0.62
    POSITIVE LOGITS
     dirty
    3.70
    dirty
    3.25
     Dirty
    2.38
     filthy
    2.00
     dirt
    1.63
     naughty
    1.60
     messy
    1.45
     clean
    1.42
     nasty
    1.37
    clean
    1.34
    Activations Density 0.022%

    No Known Activations