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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 0-RES-JB
    11. 23336
    Prev
    Next
    INDEX
    Explanations

    ingredients and cooking instructions, specifically focusing on measurements like quarts

    oai_token-act-pair · gpt-3.5-turbo

    references to measurements, particularly quantities related to recipes

    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
     Municip
    -0.74
     Cele
    -0.72
     Isle
    -0.67
     Mood
    -0.64
     Farming
    -0.63
    pic
    -0.62
    earances
    -0.61
     Folk
    -0.61
     Liv
    -0.61
     Pok
    -0.60
    POSITIVE LOGITS
    quart
    1.01
    GBT
    0.83
    >]
    0.83
    RW
    0.79
    vier
    0.78
     specialize
    0.70
    âĹ¼
    0.69
    oha
    0.69
    anson
    0.69
    ARI
    0.67
    Activations Density 0.000%

    No Known Activations