© 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 12288
    9. 8-RES_FS12288-JB
    11. 9463
    Prev
    Next
    INDEX
    Explanations

    information related to the word "well"

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    370
    +0.09
    0.4%
    261
    +0.09
    0.4%
    620
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    344
    +0.09
    0.04
    261
    +0.09
    0.04
    208
    +0.09
    0.04
    Negative Logits
    ategory
    -0.81
    hyde
    -0.79
    iferation
    -0.71
    omore
    -0.71
    rush
    -0.70
    atto
    -0.69
    iolet
    -0.67
    illary
    -0.66
    hip
    -0.66
    rely
    -0.64
    POSITIVE LOGITS
    spring
    1.24
    enough
    1.05
     behaved
    0.99
     suited
    0.94
     enough
    0.93
    esley
    0.88
     vers
    0.84
    baum
    0.83
     deserved
    0.83
    known
    0.80
    Activations Density 0.083%

    No Known Activations