© 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. 3-RES-JB
    11. 16583
    Prev
    Next
    INDEX
    Explanations

    words related to mediation and mediators

    oai_token-act-pair · gpt-3.5-turbo

    terms related to mediation and intermediaries

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    âĢ¢âĢ¢
    -0.85
    IELD
    -0.74
    ORN
    -0.71
    Leary
    -0.69
    BOOK
    -0.68
    meyer
    -0.67
     Patriot
    -0.67
    sburg
    -0.67
    âķIJ
    -0.66
    irl
    -0.66
    POSITIVE LOGITS
     medi
    1.36
    medi
    1.00
    ocr
    0.94
     mediation
    0.88
    ocre
    0.80
     negoti
    0.80
    oti
    0.77
    emate
    0.73
    ocular
    0.71
    iator
    0.70
    Activations Density 0.012%

    No Known Activations