© 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. 9259
    Prev
    Next
    INDEX
    Explanations

    mentions of politicians' names

    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
    620
    +0.12
    0.6%
    133
    +0.12
    0.5%
    592
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    232
    +0.12
    0.03
    133
    +0.12
    0.03
    659
    +0.11
    0.03
    Negative Logits
    otes
    -0.74
    683
    -0.72
    ãĥĻ
    -0.70
    ococ
    -0.69
    679
    -0.69
    690
    -0.69
    hots
    -0.68
     pse
    -0.68
     <@
    -0.66
     Scarlet
    -0.66
    POSITIVE LOGITS
    d
    1.29
    dor
    1.21
    D
    1.19
    dl
    1.14
    Ds
    1.09
    dt
    1.08
    ds
    1.07
    dL
    1.06
    dm
    1.06
    der
    1.03
    Activations Density 0.167%

    No Known Activations