© 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. 9757
    Prev
    Next
    INDEX
    Explanations

    instances of being 'included.'

    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
    558
    +0.12
    0.6%
    554
    +0.12
    0.6%
    172
    +0.12
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    558
    +0.12
    0.03
    554
    +0.12
    0.01
    498
    +0.12
    0.03
    Negative Logits
    CLASSIFIED
    -0.66
    acia
    -0.66
     Correspond
    -0.65
    ¯¯¯¯¯¯¯¯
    -0.65
    stem
    -0.62
    sis
    -0.61
    Ķ
    -0.59
    Found
    -0.59
    zee
    -0.58
    ŃĶ
    -0.58
    POSITIVE LOGITS
     prominently
    1.11
     Include
    0.83
    lihood
    0.79
    ãģĨ
    0.77
    ++++
    0.76
     herein
    0.73
     therein
    0.72
     snippets
    0.68
    clus
    0.68
    enda
    0.68
    Activations Density 0.098%

    No Known Activations