© 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. 7427
    Prev
    Next
    INDEX
    Explanations

    negative behavior or conflict in written text

    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
    +0.43
    2.1%
    627
    +0.12
    0.6%
    69
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    480
    +0.43
    0.00
    295
    +0.12
    0.05
    608
    +0.10
    0.08
    Negative Logits
     respectively
    -0.83
    )).
    -0.69
    ?".
    -0.68
    "},
    -0.68
    ?).
    -0.68
    .–
    -0.66
    %.
    -0.66
    Ħ¢
    -0.66
     etc
    -0.65
    ".[
    -0.64
    POSITIVE LOGITS
     his
    0.75
     himself
    0.71
     arra
    0.63
     voic
    0.59
     composure
    0.58
     shotgun
    0.58
     astonished
    0.57
     wheelchair
    0.55
     aback
    0.55
     1932
    0.55
    Activations Density 1.313%

    No Known Activations