© 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
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 24576
    5. 8-RES_FS24576-JB
    6. 8821
    Prev
    Next
    INDEX
    Explanations

    negative associations and societal expectations related to various topics including stigma, judgment, sexism, and body norms

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    447
    +0.15
    0.7%
    125
    +0.13
    0.6%
    40
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    724
    +0.15
    0.06
    225
    +0.13
    0.05
    125
    +0.10
    0.04
    Negative Logits
    ptroller
    -0.84
     Launcher
    -0.76
    Laun
    -0.69
    TPPStreamerBot
    -0.69
    engine
    -0.67
    issions
    -0.65
    iago
    -0.65
    hire
    -0.65
    maxwell
    -0.64
    prototype
    -0.63
    POSITIVE LOGITS
     stereotypes
    1.14
     stigmat
    1.10
     stigma
    1.10
     stereotyp
    1.10
     notions
    1.08
     dehuman
    1.05
     prejudices
    1.02
     stereotype
    1.02
     biases
    1.02
     ingrained
    1.01
    Activations Density 0.423%

    No Known Activations