© 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. 9314
    Prev
    Next
    INDEX
    Explanations

    information related to mastering in various contexts, such as music mastering services, academic degrees like a master's degree, and even hacker groups calling themselves "masters."

    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
    59
    +0.12
    0.6%
    167
    +0.12
    0.5%
    415
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    59
    +0.12
    0.02
    491
    +0.12
    0.02
    343
    +0.10
    0.02
    Negative Logits
    OPLE
    -0.78
    IDER
    -0.72
    hedon
    -0.66
    tics
    -0.65
    IGHTS
    -0.64
    enegger
    -0.63
    vernment
    -0.63
    IGH
    -0.61
    ylon
    -0.59
     SNAP
    -0.59
    POSITIVE LOGITS
    pieces
    1.49
    piece
    1.38
    mind
    1.14
    stroke
    1.03
    class
    0.91
    classes
    0.89
    abilia
    0.89
    sonian
    0.88
    work
    0.85
    fully
    0.85
    Activations Density 0.033%

    No Known Activations