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. 11-RES-JB
    6. 19752
    Prev
    Next
    INDEX
    Explanations

    phrases related to societal or political discussions, especially those discussing various viewpoints and arguments

    oai_token-act-pair · gpt-3.5-turbo

    discussions related to community issues and political discourse

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 11-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    11
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     awaited
    -0.51
     prest
    -0.51
     brink
    -0.50
     Skydragon
    -0.49
     congratulations
    -0.49
     countdown
    -0.49
    ©¶æ
    -0.48
    plete
    -0.48
    BuyableInstoreAndOnline
    -0.47
    arthed
    -0.47
    POSITIVE LOGITS
     insofar
    1.21
     when
    0.91
     whenever
    0.89
     because
    0.85
     nonetheless
    0.84
     here
    0.82
     nowadays
    0.82
    .�
    0.80
     regarding
    0.80
     if
    0.79
    Activations Density 0.565%

    No Known Activations