© 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 EvalsExportsinterp-engineNEWAPI 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. 18499
    Prev
    Next
    INDEX
    Explanations

    terms related to the acronym "CE", particularly in the context of construction and historical references

    oai_token-act-pair · gpt-3.5-turbo

    references to CE (Common Era) or related historical periods

    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
    ãĥĥãĥĪ
    -0.90
    folk
    -0.81
    eness
    -0.79
    nesota
    -0.78
    iculture
    -0.76
    reddits
    -0.74
    manager
    -0.73
    ificent
    -0.73
    gers
    -0.71
    ifying
    -0.71
    POSITIVE LOGITS
    VO
    1.08
    utical
    1.02
    VERS
    0.94
    LECT
    0.89
    lements
    0.88
    IRO
    0.87
    ASE
    0.86
    LL
    0.85
    VER
    0.84
    URRENT
    0.83
    Activations Density 0.007%

    No Known Activations