© 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
    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. 18024
    Prev
    Next
    INDEX
    Explanations

    phrases related to doubling, including actions and physical objects

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "double" and its associations with numbers, indicating a focus on doubling effects or 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
    hes
    -0.68
     Says
    -0.66
    uth
    -0.65
    âĺ
    -0.64
    politics
    -0.63
    enery
    -0.62
     Matters
    -0.61
    ankind
    -0.60
    maps
    -0.60
    ague
    -0.60
    POSITIVE LOGITS
     double
    3.57
    double
    2.76
     triple
    2.43
    Double
    2.26
     Double
    2.25
     doubles
    2.20
     quadru
    1.94
     doubled
    1.71
     doubling
    1.61
     tripled
    1.39
    Activations Density 0.019%

    No Known Activations