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. 5886
    Prev
    Next
    INDEX
    Explanations

    references to the word "Citroen"

    oai_token-act-pair · gpt-3.5-turbo

    references related to the Citroen brand or its concepts

    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
     fart
    -0.81
     helium
    -0.72
     Lovecraft
    -0.70
    leep
    -0.68
     Odin
    -0.66
     negro
    -0.66
     Hoo
    -0.66
     hem
    -0.64
    ð
    -0.64
    Mp
    -0.62
    POSITIVE LOGITS
     Barclays
    2.19
     Cit
    2.14
    TD
    1.92
     HSBC
    1.77
    CI
    1.72
     JPMorgan
    1.66
     CI
    1.60
     FD
    1.38
    td
    1.29
     TD
    1.28
    Activations Density 0.030%

    No Known Activations