© 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 768
    9. 8-RES_FS768-JB
    11. 638
    Prev
    Next
    INDEX
    Explanations

    examples of phrases discussing social or political situations and events

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs768-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    768
    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
    384
    +0.14
    0.6%
    588
    +0.10
    0.5%
    544
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    384
    +0.14
    0.11
    512
    +0.10
    0.10
    300
    +0.10
    0.14
    Negative Logits
    meet
    -0.72
     Donation
    -0.66
     Annotations
    -0.60
     Encyclopedia
    -0.58
     FANT
    -0.56
     SHARES
    -0.54
     Coup
    -0.52
     Catch
    -0.52
     ILCS
    -0.51
    ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
    -0.51
    POSITIVE LOGITS
     hers
    0.86
     ours
    0.85
     ones
    0.84
    oliath
    0.84
     theirs
    0.83
     versa
    0.81
     vice
    0.75
     outright
    0.75
     merely
    0.73
    izont
    0.69
    Activations Density 7.182%

    No Known Activations