© 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. 18865
    Prev
    Next
    INDEX
    Explanations

    mentions of various programs or initiatives

    oai_token-act-pair · gpt-3.5-turbo

    references to various types of programs

    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
    colored
    -0.88
    bors
    -0.82
     gray
    -0.82
     labeled
    -0.78
     Harbor
    -0.78
     gotten
    -0.77
     flavorful
    -0.73
     flavored
    -0.72
     flavors
    -0.72
     colored
    -0.72
    POSITIVE LOGITS
     programme
    3.90
     programmes
    3.22
     Programme
    2.83
     organisation
    1.74
     program
    1.69
     organis
    1.54
     organisers
    1.53
     organise
    1.49
     civilisation
    1.47
     analyse
    1.43
    Activations Density 0.008%

    No Known Activations