© 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 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. 6283
    Prev
    Next
    INDEX
    Explanations

    phrases related to timing such as "early", "too early", and "stage"

    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
    705
    +0.13
    0.6%
    702
    +0.11
    0.5%
    564
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    705
    +0.13
    0.05
    72
    +0.11
    0.04
    407
    +0.11
    0.03
    Negative Logits
     murderer
    -0.66
     millionaires
    -0.66
     millionaire
    -0.65
     revolt
    -0.64
     perk
    -0.61
     insult
    -0.61
    joined
    -0.60
    equality
    -0.59
     burgl
    -0.59
     quint
    -0.59
    POSITIVE LOGITS
     anecd
    0.96
     provisional
    0.96
     preliminary
    0.92
     anecdotal
    0.91
     uncertainties
    0.91
     extrap
    0.87
     TBD
    0.87
    catentry
    0.84
     unsure
    0.84
     estimates
    0.84
    Activations Density 0.428%

    No Known Activations