© 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. 9751
    Prev
    Next
    INDEX
    Explanations

    comments and code snippets related to programming errors or best practices

    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
    321
    +0.10
    0.5%
    315
    +0.10
    0.4%
    467
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.10
    0.01
    321
    +0.10
    0.02
    739
    +0.10
    0.01
    Negative Logits
    acia
    -0.79
    atre
    -0.78
    ylum
    -0.73
    agame
    -0.73
    ogy
    -0.72
    asia
    -0.70
    avorite
    -0.69
     minded
    -0.69
    eness
    -0.67
    ismo
    -0.67
    POSITIVE LOGITS
    ================================================================
    1.17
    ////////////////////////////////
    1.13
    ////////
    1.09
     ----------------------------------------------------------------
    1.08
    ERROR
    1.07
    define
    1.07
    include
    1.06
    ////////////////
    1.03
    ----------------------------------------------------------------
    1.03
    ------------------------------------------------
    1.02
    Activations Density 0.049%

    No Known Activations