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. 18051
    Prev
    Next
    INDEX
    Explanations

    email-related words and phrases

    oai_token-act-pair · gpt-3.5-turbo

    references to email communications, particularly in relation to Hillary Clinton

    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
    NT
    -0.79
     Edinburgh
    -0.78
     Rugby
    -0.77
    KR
    -0.76
     Glasgow
    -0.76
    GBT
    -0.76
     Bram
    -0.74
    axis
    -0.72
     Durham
    -0.72
     Worcester
    -0.71
    POSITIVE LOGITS
    ghazi
    1.06
     scandals
    0.87
     Emails
    0.86
     passwords
    0.85
    Clinton
    0.83
     emails
    0.83
     exchanges
    0.82
     Benghazi
    0.82
     addresses
    0.82
     dumps
    0.80
    Activations Density 0.050%

    No Known Activations