© 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. 9231
    Prev
    Next
    INDEX
    Explanations

    relationships between people and objects in a narrative context

    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
    480
    +0.26
    1.3%
    447
    +0.21
    1.1%
    246
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    480
    +0.26
    0.01
    266
    +0.21
    0.08
    720
    +0.13
    0.06
    Negative Logits
     DevOnline
    -0.77
    ĸļ
    -0.65
    estones
    -0.58
     refinement
    -0.57
    viation
    -0.56
    estone
    -0.55
    herry
    -0.53
    unciation
    -0.53
     medial
    -0.53
    isphere
    -0.52
    POSITIVE LOGITS
     THEY
    0.86
     crap
    0.80
    they
    0.79
     THEN
    0.78
     Wouldn
    0.78
    ?'
    0.77
    ?",
    0.76
     THEIR
    0.73
    ??
    0.73
     shit
    0.72
    Activations Density 0.889%

    No Known Activations