© 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. 725
    Prev
    Next
    INDEX
    Explanations

    mentions of events, releases, announcements, and declarations in a formal or official context

    oai_token-act-pair · gpt-3.5-turbo

    conjunctions and certain coordinating words in sentences

    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
    antz
    -0.80
    pher
    -0.70
    anu
    -0.70
    encies
    -0.68
    izzle
    -0.68
    onica
    -0.68
    iciency
    -0.67
    ica
    -0.66
    eness
    -0.66
    icken
    -0.65
    POSITIVE LOGITS
     Therefore
    0.77
     ãĤµãĥ¼ãĥĨãĤ£ãĥ¯ãĥ³
    0.74
     Snowden
    0.73
     QC
    0.67
     Thus
    0.67
     therefore
    0.66
     stockp
    0.66
     VER
    0.65
     Absent
    0.64
     Stalin
    0.63
    Activations Density 1.127%

    No Known Activations