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

    Neuronpedia

    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. 4-RES-JB
    11. 11387
    Prev
    Next
    INDEX
    Explanations

    mentions of the word "egg"

    oai_token-act-pair · gpt-3.5-turbo

    instances of the substring "gg"

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

    No Comments

    Negative Logits
     Spurs
    -0.66
     Hiroshima
    -0.66
     Antiqu
    -0.65
    phrine
    -0.64
     Dame
    -0.64
     Gork
    -0.64
     Mub
    -0.63
     plur
    -0.63
     Bie
    -0.63
     Declaration
    -0.61
    POSITIVE LOGITS
    ernaut
    1.08
    raphics
    1.03
    plot
    1.03
    arette
    1.01
    raft
    0.99
    iants
    0.94
    adier
    0.92
    oths
    0.92
    raphic
    0.92
    raph
    0.91
    Activations Density 0.010%

    No Known Activations