Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 4-RES-JB
    6. 3208
    Prev
    Next
    INDEX
    Explanations

    instances of the word "practice" used in a negative context

    oai_token-act-pair · gpt-3.5-turbo

    mentions of various practices, especially in legal and societal contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 4-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    4
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    onge
    -0.78
    aughter
    -0.77
    gin
    -0.76
    itialized
    -0.76
    mented
    -0.70
     raspberry
    -0.68
    zbollah
    -0.68
    aman
    -0.66
    gow
    -0.65
    aline
    -0.65
    POSITIVE LOGITS
     practiced
    0.86
     practice
    0.86
    ually
    0.84
     Practices
    0.81
    uality
    0.80
     practices
    0.78
     pract
    0.77
     Practice
    0.73
     regimen
    0.73
     practitioners
    0.73
    Activations Density 0.017%

    No Known Activations