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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW
    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    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. 11866
    Prev
    Next
    INDEX
    Explanations

    numerical quantities, especially those related to age or numbers like 30, 40, 60, etc

    oai_token-act-pair · gpt-3.5-turbo

    references to numerical values, particularly around age or 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
     https
    -0.99
    https
    -0.70
     Insp
    -0.67
    elight
    -0.66
     2017
    -0.66
     —
    -0.65
     -
    -0.65
     2016
    -0.64
     0
    -0.63
     lawmaker
    -0.63
    POSITIVE LOGITS
     twenty
    2.75
     sixty
    2.72
     thirty
    2.71
     forty
    2.70
     seventeen
    2.69
     seventy
    2.67
     fifteen
    2.66
     eighty
    2.64
     fifty
    2.64
     fourteen
    2.60
    Activations Density 0.035%

    No Known Activations