© 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. 473
    Prev
    Next
    INDEX
    Explanations

    political and social commentary

    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
    393
    +0.14
    0.6%
    622
    +0.10
    0.4%
    408
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    43
    +0.14
    0.05
    460
    +0.10
    0.04
    546
    +0.09
    0.04
    Negative Logits
    enta
    -0.88
    SourceFile
    -0.86
    usterity
    -0.80
    ighed
    -0.77
    resa
    -0.73
    onica
    -0.71
    PDATE
    -0.68
    00200000
    -0.66
    aepernick
    -0.66
    Wire
    -0.66
    POSITIVE LOGITS
     somew
    0.78
     namely
    0.60
     nil
    0.58
     acknowledges
    0.58
     depending
    0.58
     whether
    0.57
     let
    0.57
     they
    0.56
     acquies
    0.56
    there
    0.56
    Activations Density 0.342%

    No Known Activations