© 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 · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 6937
    Prev
    Next
    INDEX
    Explanations

    phrases beginning with "Think of" or "Think of it" that prompt the reader to consider a particular scenario or perspective

    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
    526
    +0.10
    0.4%
    642
    +0.10
    0.4%
    563
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    131
    +0.10
    0.01
    766
    +0.10
    0.01
    472
    +0.10
    0.01
    Negative Logits
    soever
    -0.72
    pite
    -0.68
     contended
    -0.65
    Mp
    -0.62
     contained
    -0.62
    ante
    -0.60
     subsequ
    -0.60
     declared
    -0.59
     asserted
    -0.59
     prevail
    -0.58
    POSITIVE LOGITS
     ourselves
    0.72
    agine
    0.72
     oneself
    0.70
     yourself
    0.67
     them
    0.66
    din
    0.65
    hetically
    0.64
    agus
    0.63
     ways
    0.62
     Hein
    0.61
    Activations Density 0.033%

    No Known Activations