© 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. 3-RES-JB
    11. 15486
    Prev
    Next
    INDEX
    Explanations

    phrases related to disproportionality

    oai_token-act-pair · gpt-3.5-turbo

    references to unequal or disproportionate impacts and effects within societal contexts

    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
    ince
    -0.82
    hetamine
    -0.80
    shire
    -0.79
    é¾įå
    -0.76
    hhhh
    -0.76
    ebus
    -0.75
    erm
    -0.74
    zyme
    -0.73
    uden
    -0.73
    uddled
    -0.73
    POSITIVE LOGITS
     disproportion
    0.97
     representation
    0.94
     amounts
    0.88
     consequences
    0.80
     disproportionately
    0.78
     representations
    0.78
     disadvantage
    0.77
     punishment
    0.77
     proportion
    0.76
     disadvantages
    0.76
    Activations Density 0.009%

    No Known Activations