© 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. 361
    Prev
    Next
    INDEX
    Explanations

    names of video games

    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
    104
    +0.10
    0.5%
    445
    +0.10
    0.4%
    353
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    104
    +0.10
    0.03
    424
    +0.10
    0.02
    445
    +0.10
    0.03
    Negative Logits
     âĶľâĶĢâĶĢ
    -0.67
     skim
    -0.65
    buck
    -0.62
    fusc
    -0.62
    bourg
    -0.62
    imir
    -0.61
    ©¶æ¥µ
    -0.60
     satisfies
    -0.59
     parap
    -0.58
     footing
    -0.57
    POSITIVE LOGITS
     bluff
    1.19
     Duty
    0.92
     hotline
    0.88
     shots
    0.86
     attention
    0.79
     kettle
    0.78
     reinforcements
    0.77
     timeout
    0.74
    channelAvailability
    0.73
     Haunted
    0.73
    Activations Density 0.231%

    No Known Activations