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. 0-RES-JB
    6. 3350
    Prev
    Next
    INDEX
    Explanations

    locations or organizations, particularly related to announcements or political events

    oai_token-act-pair · gpt-3.5-turbo

    occurrences of the word "Washington."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 0-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    0
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Ep
    -0.65
     PC
    -0.65
     Bj
    -0.64
     Poly
    -0.64
     Book
    -0.62
     Moor
    -0.62
     Nik
    -0.61
     inf
    -0.61
     number
    -0.61
     closed
    -0.61
    POSITIVE LOGITS
    WASHINGTON
    4.06
    ASHINGTON
    2.12
    Washington
    1.90
    SAN
    1.53
    NEW
    1.36
    TRUMP
    1.35
    SHARE
    1.35
     YORK
    1.32
    Congress
    1.28
    LOS
    1.28
    Activations Density 0.017%

    No Known Activations