© 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. 20626
    Prev
    Next
    INDEX
    Explanations

    information or situations that are recurring or done consistently as a practice by various entities

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of the word "routinely."

    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
    lad
    -0.66
     Straw
    -0.66
     Millennium
    -0.65
     Emir
    -0.64
     Chains
    -0.62
    iant
    -0.61
     Redemption
    -0.60
     Ruler
    -0.60
     Tycoon
    -0.60
    busters
    -0.60
    POSITIVE LOGITS
    theless
    0.90
    entimes
    0.83
     pract
    0.82
     disreg
    0.82
    ãĤ©
    0.80
     subjected
    0.75
     accommod
    0.74
     behaved
    0.74
     ignored
    0.74
     employed
    0.74
    Activations Density 0.014%

    No Known Activations