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. 0-RES-JB
    6. 8571
    Prev
    Next
    INDEX
    Explanations

    instances where an amount or quantity has doubled

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "doubled" or related terms indicating an increase in quantity

    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
    bm
    -0.66
    view
    -0.65
     Struggle
    -0.62
    il
    -0.62
    Nav
    -0.62
    atu
    -0.61
    forums
    -0.60
     Mich
    -0.59
    api
    -0.58
    topic
    -0.58
    POSITIVE LOGITS
     doubled
    3.43
     tripled
    2.98
     doubling
    2.67
     quadru
    2.21
     doubles
    2.09
     double
    1.56
     hal
    1.31
     increased
    1.25
     triple
    1.25
    double
    1.24
    Activations Density 0.013%

    No Known Activations