© 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. 12743
    Prev
    Next
    INDEX
    Explanations

    statements that include comparisons or evaluations such as "thinking about," "considering," and "looking at."

    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
    373
    +0.18
    0.8%
    429
    +0.12
    0.6%
    218
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    715
    +0.18
    0.04
    750
    +0.12
    0.03
    157
    +0.11
    0.04
    Negative Logits
    arsity
    -0.75
    Els
    -0.72
    burgh
    -0.67
    Es
    -0.66
    ãĤ´
    -0.66
    Requires
    -0.65
    Guest
    -0.63
    phrine
    -0.61
    esm
    -0.61
    ERN
    -0.60
    POSITIVE LOGITS
     hindsight
    0.80
    »Ĵ
    0.79
     extrap
    0.78
     hypot
    0.77
     retrospect
    0.77
    Compare
    0.77
     analogy
    0.76
     comparing
    0.74
     perspective
    0.74
     context
    0.72
    Activations Density 0.159%

    No Known Activations