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. 21473
    Prev
    Next
    INDEX
    Explanations

    company or place names related to business ventures

    oai_token-act-pair · gpt-3.5-turbo

    references to specific entities or people, particularly related to the names Fen and Bain, along with the term "dealings."

    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
    tails
    -0.87
     drag
    -0.68
     POP
    -0.66
     prelim
    -0.66
    asses
    -0.64
    2018
    -0.64
     Wag
    -0.63
    umbo
    -0.63
    oby
    -0.62
    osc
    -0.61
    POSITIVE LOGITS
     Fen
    2.17
     Bain
    2.13
     Anfield
    1.37
     Eden
    1.31
     dealings
    1.20
     blight
    1.18
     Beacon
    1.10
     stature
    1.01
     Gerrard
    1.01
     Stamford
    1.01
    Activations Density 0.044%

    No Known Activations