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. 3-RES-JB
    6. 19844
    Prev
    Next
    INDEX
    Explanations

    phrases related to the United States political system and government

    oai_token-act-pair · gpt-3.5-turbo

    references to the United States government and its institutions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    hots
    -0.85
    develop
    -0.77
    DragonMagazine
    -0.74
    fig
    -0.68
    alys
    -0.68
    BUT
    -0.68
     livest
    -0.67
    PsyNetMessage
    -0.64
    figure
    -0.63
    merce
    -0.63
    POSITIVE LOGITS
     Constitution
    0.95
     Treasury
    0.94
     Citizenship
    0.88
     constitutionally
    0.81
     Accountability
    0.81
     electorate
    0.80
     election
    0.78
     constitutional
    0.76
     defund
    0.76
     treasury
    0.75
    Activations Density 0.185%

    No Known Activations