© 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 49152
    9. 8-RES_FS49152-JB
    11. 26602
    Prev
    Next
    INDEX
    Explanations

    numerals and terms related to measurements or academic metrics

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    285
    +0.14
    0.6%
    195
    +0.13
    0.6%
    738
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    325
    +0.14
    0.01
    285
    +0.13
    0.01
    592
    +0.12
    0.01
    Negative Logits
    zynski
    -0.93
    andowski
    -0.90
    clip
    -0.83
    horn
    -0.81
    doms
    -0.79
    idem
    -0.78
    clips
    -0.77
    gren
    -0.76
    helm
    -0.72
    backs
    -0.71
    POSITIVE LOGITS
     pH
    1.02
     gradient
    0.76
    BG
    0.72
    #$
    0.72
     sensitivity
    0.71
     acidic
    0.70
     readings
    0.70
     Calculator
    0.69
     calculator
    0.67
     variance
    0.66
    Activations Density 0.008%

    No Known Activations