© 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 · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 24576
    9. 8-RES_FS24576-JB
    11. 4798
    Prev
    Next
    INDEX
    Explanations

    phrases related to introspection and self-reflection

    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
    608
    +0.14
    0.7%
    29
    +0.11
    0.5%
    7
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    608
    +0.14
    0.06
    22
    +0.11
    0.05
    123
    +0.11
    0.05
    Negative Logits
    NOW
    -0.68
    aunder
    -0.64
    llor
    -0.62
    linger
    -0.60
    EVA
    -0.60
    BUS
    -0.60
    book
    -0.59
    killed
    -0.59
    IDE
    -0.59
    pedia
    -0.58
    POSITIVE LOGITS
    effic
    1.43
     lieu
    1.29
     spite
    1.20
    ordinate
    1.18
    escap
    1.17
    efficiency
    1.17
     relation
    1.11
     accordance
    1.11
     regards
    1.02
    animate
    1.01
    Activations Density 0.206%

    No Known Activations