© 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. 8-RES-JB
    11. 6143
    Prev
    Next
    INDEX
    Explanations

    phrases related to physical injuries and medical procedures

    oai_token-act-pair · gpt-3.5-turbo

    conjunctions, particularly the word "and."

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    Redditor
    -0.75
     bumped
    -0.64
    rb
    -0.64
    `
    -0.63
    ´
    -0.63
    aldo
    -0.63
    js
    -0.62
    jar
    -0.62
     Drawn
    -0.62
    :]
    -0.62
    POSITIVE LOGITS
     assorted
    1.11
     possibly
    0.87
     downright
    0.75
    etc
    0.73
    whatever
    0.72
     other
    0.71
    ospace
    0.69
     even
    0.69
     ultimately
    0.66
     others
    0.66
    Activations Density 0.162%

    No Known Activations