© 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 12288
    9. 8-RES_FS12288-JB
    11. 404
    Prev
    Next
    INDEX
    Explanations

    keywords related to taxes, including "capital gains."

    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
    288
    +0.19
    0.9%
    459
    +0.11
    0.5%
    247
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    288
    +0.19
    0.03
    247
    +0.11
    0.02
    404
    +0.11
    0.02
    Negative Logits
    £ı
    -0.88
    ezvous
    -0.78
    ĪĴ
    -0.77
    DEN
    -0.77
     Schr
    -0.71
    MAL
    -0.67
     Shack
    -0.67
    SEE
    -0.67
    ya
    -0.67
    ller
    -0.66
    POSITIVE LOGITS
    ization
    1.02
    istic
    0.98
    izations
    0.98
    izes
    0.95
    ized
    0.94
    izing
    0.92
    orpor
    0.91
     accumulation
    0.91
     punishment
    0.91
    isation
    0.88
    Activations Density 0.040%

    No Known Activations