© 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. 4903
    Prev
    Next
    INDEX
    Explanations

    words related to technological advancements and developments

    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
    762
    +0.11
    0.5%
    566
    +0.11
    0.5%
    545
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    344
    +0.11
    0.08
    668
    +0.11
    0.06
    566
    +0.10
    0.07
    Negative Logits
     Apart
    -0.61
     Phar
    -0.61
     Anyway
    -0.60
     Useful
    -0.57
     Carth
    -0.55
     panel
    -0.55
     Worth
    -0.55
     Plymouth
    -0.54
     Nielsen
    -0.54
     Pair
    -0.53
    POSITIVE LOGITS
     been
    1.36
     become
    1.20
    been
    1.13
     gotten
    1.07
     undergone
    1.05
     struggled
    1.03
     endeav
    1.01
     begun
    0.98
     risen
    0.96
     gone
    0.94
    Activations Density 0.250%

    No Known Activations