© 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. 22281
    Prev
    Next
    INDEX
    Explanations

    large quantities or numbers

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "thousand" and its variations in the text

    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
    appa
    -0.61
    sd
    -0.59
     Hook
    -0.59
    PD
    -0.58
    News
    -0.58
     Expand
    -0.57
    emouth
    -0.56
    Bench
    -0.56
     intro
    -0.56
    kell
    -0.55
    POSITIVE LOGITS
     thousand
    3.49
     hundred
    2.62
     Thousand
    2.16
    ousand
    1.80
     million
    1.80
     dozen
    1.77
     lakh
    1.64
     thousands
    1.60
     Hundred
    1.58
     thous
    1.51
    Activations Density 0.016%

    No Known Activations