© 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. 0-RES-JB
    11. 13604
    Prev
    Next
    INDEX
    Explanations

    past tense verbs, specifically those ending in 'ed'

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "did."

    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
    clusively
    -0.58
    alg
    -0.57
    ILY
    -0.56
    immer
    -0.56
    agonists
    -0.54
    anchester
    -0.52
     Flavoring
    -0.51
    Merit
    -0.51
    room
    -0.49
    immers
    -0.49
    POSITIVE LOGITS
     did
    2.88
    did
    2.34
     DID
    1.91
     didn
    1.85
     does
    1.84
     didnt
    1.64
     Did
    1.64
     undertook
    1.57
    didn
    1.56
    Did
    1.56
    Activations Density 0.049%

    No Known Activations