© 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. 1-RES-JB
    11. 2291
    Prev
    Next
    INDEX
    Explanations

    words related to uncleanliness or unconventional methods

    oai_token-act-pair · gpt-3.5-turbo

    references to the term "dirty," indicating a focus on negative or undesirable conditions or actions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 1-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    1
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    HCR
    -0.99
    iphate
    -0.92
    mare
    -0.86
    */(
    -0.81
    zai
    -0.77
    hari
    -0.76
    âĹ¼
    -0.76
    ItemThumbnailImage
    -0.75
    ommel
    -0.75
    ĸļ
    -0.75
    POSITIVE LOGITS
     dirty
    1.12
     laundry
    1.05
    dirty
    0.96
     linen
    0.88
     mole
    0.85
     cleaner
    0.85
     Dirty
    0.81
     Clean
    0.79
     dirt
    0.77
    clean
    0.77
    Activations Density 0.008%

    No Known Activations