© 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 EvalsExportsinterp-engineNEWAPI 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. 15354
    Prev
    Next
    INDEX
    Explanations

    words related to physical touch and affection, specifically 'cuddling'

    oai_token-act-pair · gpt-3.5-turbo

    variations of the word "uddle," indicating an interest in concepts related to clustering or gathering

    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
     production
    -0.68
     Prairie
    -0.67
     credit
    -0.67
     operations
    -0.65
     servers
    -0.65
     AGA
    -0.65
     terminal
    -0.64
     expect
    -0.64
     gate
    -0.63
     Colon
    -0.63
    POSITIVE LOGITS
    udd
    4.48
    uddled
    2.11
    uddle
    2.03
    uddy
    1.61
    udding
    1.50
    uds
    1.27
    uzz
    1.20
    ud
    1.14
    udder
    1.13
    idd
    1.09
    Activations Density 0.020%

    No Known Activations