© 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. 3-RES-JB
    11. 7230
    Prev
    Next
    INDEX
    Explanations

    instances of the word "special" with high relevance

    oai_token-act-pair · gpt-3.5-turbo

    the word "special" in various contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 3-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    3
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Twain
    -0.77
     Couch
    -0.73
     Ri
    -0.69
     Yose
    -0.66
     Giul
    -0.66
     Webster
    -0.64
     Merrill
    -0.63
     Neighbor
    -0.63
     Cah
    -0.62
     Springer
    -0.61
    POSITIVE LOGITS
    ised
    1.15
    ties
    1.07
    isations
    1.06
    izations
    0.95
    isable
    0.94
    marine
    0.92
    ombat
    0.91
    nels
    0.84
    ises
    0.82
    atural
    0.82
    Activations Density 0.018%

    No Known Activations