© 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. 10-RES-JB
    11. 5913
    Prev
    Next
    INDEX
    Explanations

    contractions of words like "will not" into "won't"

    oai_token-act-pair · gpt-3.5-turbo

    repeated negations of the word "won't."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 10-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    10
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    anwhile
    -0.68
     background
    -0.65
     beginnings
    -0.65
     Background
    -0.65
     backdrop
    -0.64
     populated
    -0.63
     seeded
    -0.61
     watered
    -0.60
     newcom
    -0.60
     sket
    -0.59
    POSITIVE LOGITS
    't
    2.04
    kish
    1.18
    cest
    1.14
    now
    1.10
    itive
    1.01
    ky
    1.00
    ced
    0.98
    ´
    0.98
    cing
    0.97
    throp
    0.94
    Activations Density 0.041%

    No Known Activations