© 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. 19992
    Prev
    Next
    INDEX
    Explanations

    references to the meal "dinner" being prepared or consumed

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "dinner" and related contexts

    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
     Manip
    -0.68
    olan
    -0.67
    ases
    -0.66
     Morph
    -0.65
    ITH
    -0.65
    uitive
    -0.64
     Areas
    -0.63
    cent
    -0.60
    currently
    -0.59
    forces
    -0.59
    POSITIVE LOGITS
     dinner
    3.55
     dinners
    2.75
     supper
    2.74
     Dinner
    2.71
     lunch
    2.46
     breakfast
    2.38
     meal
    2.03
     brunch
    2.02
     meals
    1.94
     dessert
    1.75
    Activations Density 0.023%

    No Known Activations