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. 7122
    Prev
    Next
    INDEX
    Explanations

    numbers, specifically the digit "Thousand" represented by the number 10

    oai_token-act-pair · gpt-3.5-turbo

    numeric representations of large quantities

    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
     Hook
    -0.69
     Record
    -0.69
     Play
    -0.68
     Countdown
    -0.66
    oke
    -0.65
     Story
    -0.64
    rx
    -0.63
     Bloomberg
    -0.63
     Simulator
    -0.62
     Motors
    -0.62
    POSITIVE LOGITS
     Thousand
    1.24
     Cth
    1.00
    zhen
    0.91
     redes
    0.84
    antha
    0.80
     iceberg
    0.76
    istani
    0.71
     oneself
    0.69
    Balt
    0.65
    appa
    0.65
    Activations Density 0.000%

    No Known Activations