© 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. 9726
    Prev
    Next
    INDEX
    Explanations

    headlines or headings in a text

    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
    159
    +0.12
    0.6%
    394
    +0.12
    0.5%
    609
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    609
    +0.12
    0.02
    281
    +0.12
    0.01
    288
    +0.11
    0.02
    Negative Logits
    ajor
    -0.76
    enhagen
    -0.76
    ikarp
    -0.75
    sterdam
    -0.74
    ono
    -0.74
    cible
    -0.73
    ederation
    -0.71
    ello
    -0.71
    atron
    -0.71
    aum
    -0.69
    POSITIVE LOGITS
     headlines
    1.07
     headline
    1.05
     proclaiming
    0.91
    lined
    0.86
     headlined
    0.77
    ãĥŀ
    0.75
    ocl
    0.73
    vier
    0.69
    dom
    0.68
    tenance
    0.67
    Activations Density 0.090%

    No Known Activations