© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    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. 3-RES-JB
    11. 3048
    Prev
    Next
    INDEX
    Explanations

    contractions of the words "could" and "not"

    oai_token-act-pair · gpt-3.5-turbo

    the phrase "couldn't" in various contexts

    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
    oak
    -0.74
    dress
    -0.73
    backer
    -0.72
     Mant
    -0.68
    bred
    -0.65
     PU
    -0.63
     Rising
    -0.60
     GB
    -0.59
     duty
    -0.59
    chel
    -0.58
    POSITIVE LOGITS
    't
    1.55
    adian
    0.95
     afford
    0.89
     feas
    0.86
    atio
    0.82
    ieve
    0.82
    berra
    0.79
    ayan
    0.79
    itial
    0.77
    ÃĥÃĤ
    0.77
    Activations Density 0.013%

    No Known Activations