© 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. 45461
    Prev
    Next
    INDEX
    Explanations

    references to the concept of being "halfway" or "midway" in various contexts

    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
    339
    +0.12
    0.6%
    48
    +0.10
    0.4%
    333
    +0.08
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    339
    +0.12
    0.01
    333
    +0.10
    0.01
    190
    +0.08
    0.01
    Negative Logits
    ulators
    -0.86
    asts
    -0.79
    UL
    -0.78
    ULE
    -0.76
    rum
    -0.72
    Trend
    -0.70
    rums
    -0.69
    oran
    -0.67
    oppable
    -0.67
    eer
    -0.66
    POSITIVE LOGITS
     halfway
    0.88
     ILCS
    0.73
    wegian
    0.72
    assed
    0.68
     toward
    0.66
     decent
    0.66
    heartedly
    0.65
    borg
    0.65
    terness
    0.64
    elfth
    0.64
    Activations Density 0.005%

    No Known Activations