© 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 24576
    9. 8-RES_FS24576-JB
    11. 18832
    Prev
    Next
    INDEX
    Explanations

    terms related to comparison and contrast

    oai_token-act-pair · gpt-3.5-turboTriggered 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
    480
    +0.18
    0.8%
    447
    +0.16
    0.7%
    511
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    511
    +0.18
    0.03
    632
    +0.16
    0.03
    441
    +0.10
    0.03
    Negative Logits
    Ire
    -0.66
    athan
    -0.62
    osc
    -0.57
    ãģł
    -0.56
    eur
    -0.56
    Redditor
    -0.56
    åħ
    -0.55
    å½
    -0.55
    å£
    -0.55
    Category
    -0.54
    POSITIVE LOGITS
    pires
    1.15
    pire
    1.10
    pired
    1.07
    cription
    1.06
    ynchron
    1.04
     opposed
    1.00
    bestos
    1.00
    semb
    0.96
    pects
    0.96
    phy
    0.95
    Activations Density 0.054%

    No Known Activations