© 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 24576
    9. 8-RES_FS24576-JB
    11. 24279
    Prev
    Next
    INDEX
    Explanations

    descriptive adjectives related to various characteristics and qualities

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    447
    +0.13
    0.6%
    244
    +0.12
    0.5%
    70
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    271
    +0.13
    0.06
    70
    +0.12
    0.06
    588
    +0.10
    0.06
    Negative Logits
    ourt
    -0.69
    rez
    -0.66
    tein
    -0.63
    onne
    -0.59
    ribe
    -0.58
    annabin
    -0.58
    ocations
    -0.58
    ortmund
    -0.57
     largeDownload
    -0.57
    secution
    -0.57
    POSITIVE LOGITS
     enough
    1.67
    enough
    1.07
     Enough
    1.06
     insofar
    0.87
     nonetheless
    0.85
    wired
    0.80
     because
    0.78
    istically
    0.77
     surpr
    0.77
     beyond
    0.77
    Activations Density 0.339%

    No Known Activations