© 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. 2131
    Prev
    Next
    INDEX
    Explanations

    words related to physical structures or positions, like "fronts," "facade," and "storefronts."

    oai_token-act-pair · gpt-3.5-turbo

    references to various aspects of design and structural elements related to buildings or settings

    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
    ciating
    -0.95
    ascript
    -0.78
    ilon
    -0.77
    omial
    -0.73
    ioxide
    -0.71
    ilar
    -0.69
    otos
    -0.69
    ipop
    -0.67
    olved
    -0.66
    animous
    -0.66
    POSITIVE LOGITS
     fronts
    0.98
     facade
    0.83
     storefront
    0.80
     Dres
    0.73
    soDeliveryDate
    0.72
     wip
    0.70
    terness
    0.68
    RPG
    0.66
    isSpecialOrderable
    0.64
    FML
    0.62
    Activations Density 0.012%

    No Known Activations