© 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. 37445
    Prev
    Next
    INDEX
    Explanations

    phrases that indicate contrasts between different states or conditions

    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
    481
    +0.17
    0.8%
    517
    +0.12
    0.5%
    669
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    669
    +0.17
    0.03
    362
    +0.12
    0.04
    608
    +0.10
    0.03
    Negative Logits
    Solution
    -0.79
    acha
    -0.75
    etimes
    -0.72
    lihood
    -0.68
     avail
    -0.65
    CVE
    -0.65
    can
    -0.64
     succeeded
    -0.64
    ãĥ¯
    -0.63
    amen
    -0.63
    POSITIVE LOGITS
     precon
    0.98
     shack
    0.94
     dependence
    0.92
     clutter
    0.92
     stereotypes
    0.92
     negativity
    0.90
     confines
    0.90
     reliance
    0.88
     fear
    0.86
     distractions
    0.85
    Activations Density 0.112%

    No Known Activations