© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 37581
    Prev
    Next
    INDEX
    Explanations

    conjunctions and phrases that indicate connections or relationships between ideas

    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
    +0.12
    0.5%
    284
    +0.11
    0.5%
    537
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    284
    +0.12
    0.04
    393
    +0.11
    0.04
    384
    +0.11
    0.04
    Negative Logits
    wic
    -0.77
    -------
    -0.73
    anmar
    -0.73
    interrupted
    -0.72
    sters
    -0.72
    Camp
    -0.71
    MET
    -0.71
    involved
    -0.71
    cffff
    -0.71
    oused
    -0.70
    POSITIVE LOGITS
     decency
    0.88
     prestige
    0.85
     architectural
    0.84
     guiActiveUnfocused
    0.84
     ideals
    0.84
     likeness
    0.83
     significance
    0.82
     purity
    0.82
     usefulness
    0.80
     individuality
    0.80
    Activations Density 0.155%

    No Known Activations