© 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. 0-RES-JB
    11. 22517
    Prev
    Next
    INDEX
    Explanations

    references to the United States (U.S.)

    oai_token-act-pair · gpt-3.5-turbo

    the letter 'S' at the beginning of various phrases and terms

    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
     derog
    -0.63
     balloon
    -0.61
     advis
    -0.56
     unavoidable
    -0.56
     vetting
    -0.56
     elbow
    -0.54
     mild
    -0.54
     homework
    -0.54
     gru
    -0.54
     hoop
    -0.54
    POSITIVE LOGITS
    S
    3.32
    SG
    1.60
    Si
    1.53
     S
    1.51
    Cs
    1.47
    SR
    1.40
    C
    1.37
    s
    1.36
    R
    1.35
    P
    1.32
    Activations Density 0.073%

    No Known Activations