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. 7-RES-JB
    6. 18292
    Prev
    Next
    INDEX
    Explanations

    phrases containing the word "of" followed by a specific entity or concept

    oai_token-act-pair · gpt-3.5-turbo

    phrases related to government agencies and departments

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 7-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    7
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    NetMessage
    -0.81
     arteries
    -0.74
    wcs
    -0.73
     hots
    -0.70
    comings
    -0.68
     hinges
    -0.67
     corro
    -0.64
    reddits
    -0.64
    rings
    -0.64
     pots
    -0.62
    POSITIVE LOGITS
     Agriculture
    1.03
     Commerce
    1.02
     Transportation
    1.01
     Justice
    0.96
     Education
    0.90
     Corrections
    0.89
     Energy
    0.87
     Revenue
    0.86
     Health
    0.85
     Communications
    0.85
    Activations Density 0.041%

    No Known Activations