© 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. 22823
    Prev
    Next
    INDEX
    Explanations

    phrases indicating speculation or uncertainty about knowledge or experiences

    oai_token-act-pair · gpt-4o-miniTriggered 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
    373
    +0.11
    0.5%
    726
    +0.09
    0.4%
    174
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    566
    +0.11
    0.04
    763
    +0.09
    0.03
    583
    +0.09
    0.03
    Negative Logits
     unification
    -0.66
     openness
    -0.65
     merger
    -0.62
     Dodd
    -0.60
     Instit
    -0.58
     cardinal
    -0.58
     fundament
    -0.58
     sacrific
    -0.57
     negoti
    -0.57
     Fal
    -0.57
    POSITIVE LOGITS
     noticed
    1.17
     guessed
    1.07
     wondering
    1.05
    swer
    0.96
     noticing
    0.94
     yourselves
    0.93
     wondered
    0.92
     hearing
    0.90
     WATCHED
    0.90
     reading
    0.88
    Activations Density 0.193%

    No Known Activations