© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 1-RES-JB
    11. 22017
    Prev
    Next
    INDEX
    Explanations

    mentions of the brand "Coca-Cola" and related terms

    oai_token-act-pair · gpt-3.5-turbo

    mentions of brand names, specifically Coca-Cola and FIA

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 1-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    1
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    */(
    -0.71
    Redditor
    -0.70
    yrim
    -0.69
    oppable
    -0.67
    ORN
    -0.67
    ebook
    -0.67
    ulo
    -0.67
    atche
    -0.66
     [+
    -0.64
    Rog
    -0.64
    POSITIVE LOGITS
     Coca
    0.89
    cedes
    0.87
     lapt
    0.87
    Cola
    0.79
     Coke
    0.78
    gow
    0.78
    cers
    0.77
     nutrit
    0.77
     encour
    0.75
    ciating
    0.74
    Activations Density 0.007%

    No Known Activations