© 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. 13865
    Prev
    Next
    INDEX
    Explanations

    word patterns related to software programming languages and technology concepts, particularly those involving "Bayesian", "C++", and "programmer"

    oai_token-act-pair · gpt-3.5-turbo

    references to a specific geographical location, named entities, particularly related to "Bay" and "Bike."

    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
     char
    -0.69
     forces
    -0.67
     effects
    -0.64
     clarity
    -0.64
     permission
    -0.64
     solemn
    -0.63
     consequences
    -0.63
     indic
    -0.63
     grit
    -0.62
     interruption
    -0.62
    POSITIVE LOGITS
    Bay
    4.04
    bay
    2.53
    bike
    1.31
    Ju
    1.09
    bath
    1.09
    insula
    0.98
    sonian
    0.98
    Ber
    0.97
    bern
    0.96
    commerce
    0.96
    Activations Density 0.013%

    No Known Activations