© 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. 4-RES-JB
    11. 6944
    Prev
    Next
    INDEX
    Explanations

    phrases related to specific locations or entities spelled in a particular way

    oai_token-act-pair · gpt-3.5-turbo

    terms related to administrative divisions or geographical entities

    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
    ppo
    -0.71
     Forge
    -0.70
    bians
    -0.70
    ulkan
    -0.69
    tes
    -0.69
    ppa
    -0.68
    hene
    -0.67
    kies
    -0.66
    zo
    -0.66
    izations
    -0.66
    POSITIVE LOGITS
    irect
    1.08
    ead
    1.02
    aniel
    1.00
    ocument
    1.00
    irection
    0.98
    ragon
    0.95
    ynamic
    0.95
    ouble
    0.94
    isd
    0.94
    itional
    0.88
    Activations Density 0.030%

    No Known Activations