© 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. 4-RES-JB
    11. 1019
    Prev
    Next
    INDEX
    Explanations

    words related to financial matters and economic activities

    oai_token-act-pair · gpt-3.5-turbo

    informal slang terms or abbreviations

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

    No Comments

    Negative Logits
    hyde
    -0.80
    DragonMagazine
    -0.75
     medium
    -0.72
     Phant
    -0.65
    ocene
    -0.65
     Sorceress
    -0.63
     sacrament
    -0.63
     Norse
    -0.63
    Origin
    -0.61
    adobe
    -0.61
    POSITIVE LOGITS
    ashing
    0.98
    ailing
    0.92
    ashed
    0.91
    inged
    0.90
    uss
    0.87
    inging
    0.87
    ibr
    0.87
    urring
    0.86
    aff
    0.86
    icating
    0.84
    Activations Density 0.287%

    No Known Activations