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

    Neuronpedia

    Jacobian LensNEW
    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. 3822
    Prev
    Next
    INDEX
    Explanations

    words related to strong negative emotions or physical suffering

    oai_token-act-pair · gpt-3.5-turbo

    references to various forms of distress

    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
    hetti
    -0.71
     Swords
    -0.70
    tein
    -0.70
    bern
    -0.65
    stood
    -0.63
    ioch
    -0.63
    glas
    -0.62
    ramid
    -0.62
    zyme
    -0.62
    llan
    -0.61
    POSITIVE LOGITS
     distress
    1.17
    ingly
    1.07
    rained
    0.73
     distressed
    0.72
    fully
    0.71
     Jagu
    0.68
    ibly
    0.68
    chenko
    0.68
     respons
    0.66
    OUS
    0.66
    Activations Density 0.009%

    No Known Activations