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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports 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. 9484
    Prev
    Next
    INDEX
    Explanations

    verbs expressing possibility or expectation

    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
    439
    +0.10
    0.5%
    429
    +0.10
    0.5%
    373
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    551
    +0.10
    0.03
    718
    +0.10
    0.03
    429
    +0.10
    0.02
    Negative Logits
    DERR
    -0.68
     percussion
    -0.68
     Countdown
    -0.65
    Enough
    -0.64
     Bots
    -0.63
     Daddy
    -0.63
    Cook
    -0.62
    GGGGGGGG
    -0.61
    CG
    -0.59
     drums
    -0.59
    POSITIVE LOGITS
     speculate
    1.05
     wonder
    1.03
     imagine
    0.99
     argue
    0.95
     presume
    0.93
     assume
    0.92
    £ı
    0.92
     wonders
    0.91
     ponder
    0.90
     conclude
    0.88
    Activations Density 0.092%

    No Known Activations