© 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. 17205
    Prev
    Next
    INDEX
    Explanations

    phrases related to the assessment or critique of processes or events

    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
    +0.12
    0.5%
    304
    +0.10
    0.5%
    512
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    512
    +0.12
    0.06
    36
    +0.10
    0.06
    304
    +0.10
    0.05
    Negative Logits
    existent
    -0.74
    styles
    -0.70
    english
    -0.69
     Languages
    -0.69
     bathrooms
    -0.69
     prefers
    -0.68
     disappears
    -0.68
     liking
    -0.68
     toilets
    -0.67
     varies
    -0.66
    POSITIVE LOGITS
     culmination
    1.20
     testament
    0.98
     acknowledgment
    0.96
     vind
    0.95
     acknowledgement
    0.94
     reaff
    0.92
     unprecedented
    0.91
     validation
    0.89
     reminder
    0.89
     continuation
    0.87
    Activations Density 0.379%

    No Known Activations