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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 7987
    Prev
    Next
    INDEX
    Explanations

    the word "debut" in various contexts

    oai_token-act-pair · gpt-3.5-turbo

    instances of the term "debate."

    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
    pecting
    -0.73
    lift
    -0.73
    ĸļ
    -0.70
    oreal
    -0.69
    itivity
    -0.66
    Catalog
    -0.65
    QL
    -0.65
    onics
    -0.63
    oop
    -0.63
    ciating
    -0.62
    POSITIVE LOGITS
     deb
    4.13
    deb
    2.28
    Deb
    1.95
     Deb
    1.85
     dep
    1.14
     premie
    1.07
     buffs
    1.00
     buff
    0.98
     flo
    0.98
     Dise
    0.97
    Activations Density 0.025%

    No Known Activations