© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    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. 22153
    Prev
    Next
    INDEX
    Explanations

    conjunctions and adverbs that indicate contrast or exceptions

    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
    +0.14
    0.6%
    544
    +0.11
    0.5%
    756
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    546
    +0.14
    0.08
    43
    +0.11
    0.07
    418
    +0.09
    0.06
    Negative Logits
    iple
    -0.62
    itol
    -0.61
    ©¶æ
    -0.60
    iman
    -0.56
    ivity
    -0.56
    ĨĴ
    -0.55
     è£ıè
    -0.53
     resemb
    -0.53
     Lis
    -0.53
    isode
    -0.53
    POSITIVE LOGITS
     respectively
    1.35
     whichever
    0.98
    depending
    0.86
     depending
    0.84
     albeit
    0.80
     Including
    0.70
     haha
    0.70
     alike
    0.69
     although
    0.69
     accordingly
    0.69
    Activations Density 0.513%

    No Known Activations