© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 93
    Prev
    Next
    INDEX
    Explanations

    phrases related to politics and geopolitical issues

    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
    323
    +0.10
    0.4%
    95
    +0.09
    0.4%
    376
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    294
    +0.10
    0.03
    45
    +0.09
    0.03
    393
    +0.08
    0.07
    Negative Logits
     sugg
    -0.88
     impair
    -0.76
     behaviors
    -0.75
    requently
    -0.73
     impairment
    -0.70
     behaviours
    -0.68
     impaired
    -0.67
     hypothesized
    -0.66
     behavioral
    -0.64
     outlined
    -0.64
    POSITIVE LOGITS
     blah
    1.12
    HAHAHAHA
    0.80
    indal
    0.76
    xxx
    0.76
     etc
    0.75
     Nope
    0.74
    Ü
    0.72
     yeah
    0.71
     Columb
    0.71
    xxxx
    0.70
    Activations Density 0.481%

    No Known Activations