Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 11-RES-JB
    6. 18256
    Prev
    Next
    INDEX
    Explanations

    occurrences of the word "but" with a higher activation value than other tokens

    oai_token-act-pair · gpt-3.5-turbo

    the conjunction "but" indicating contrasts or exceptions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Cycle
    -0.68
     Excellence
    -0.62
    naire
    -0.55
     Improvement
    -0.54
     Procedures
    -0.53
    ĺħ
    -0.53
    ampa
    -0.53
     stimulus
    -0.52
    ESH
    -0.52
     pursu
    -0.51
    POSITIVE LOGITS
    chers
    1.39
    chery
    1.28
    tons
    1.14
    ts
    1.04
     alas
    0.97
    ted
    0.96
     nevertheless
    0.92
    ters
    0.92
     nonetheless
    0.90
    cher
    0.90
    Activations Density 0.089%

    No Known Activations