© 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. 303
    Prev
    Next
    INDEX
    Explanations

    information related to criminal court cases

    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
    480
    +0.41
    2.0%
    420
    +0.10
    0.5%
    79
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    480
    +0.41
    -0.01
    86
    +0.10
    0.07
    266
    +0.10
    0.14
    Negative Logits
     favors
    -0.90
    defense
    -0.88
    avored
    -0.87
    Defense
    -0.87
    colored
    -0.86
     favored
    -0.86
    imize
    -0.85
     someday
    -0.83
     favorably
    -0.82
     favor
    -0.81
    POSITIVE LOGITS
     £
    1.55
     recognise
    1.44
     realise
    1.41
     apologised
    1.37
     centre
    1.36
     Scotland
    1.35
     apologise
    1.34
     Yorkshire
    1.34
     Edinburgh
    1.33
     recognised
    1.32
    Activations Density 3.251%

    No Known Activations