© 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 12288
    9. 8-RES_FS12288-JB
    11. 9629
    Prev
    Next
    INDEX
    Explanations

    proper nouns and names

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    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
    87
    +0.17
    0.9%
    712
    +0.14
    0.7%
    324
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    232
    +0.17
    0.10
    659
    +0.14
    0.09
    145
    +0.10
    0.07
    Negative Logits
     silence
    -0.58
     apologies
    -0.58
     pleas
    -0.58
     hindsight
    -0.57
     gamb
    -0.57
     peer
    -0.56
     abst
    -0.56
     econom
    -0.56
     impulse
    -0.55
     mend
    -0.55
    POSITIVE LOGITS
    ensis
    0.93
    orum
    0.81
     Remix
    0.78
     LLP
    0.76
     Ltd
    0.76
     Edition
    0.76
    astery
    0.69
    nr
    0.68
    naire
    0.67
     LTD
    0.67
    Activations Density 1.204%

    No Known Activations