© 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 49152
    9. 8-RES_FS49152-JB
    11. 37389
    Prev
    Next
    INDEX
    Explanations

    verbs related to actions performed or processes carried out

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    756
    +0.10
    0.5%
    126
    +0.10
    0.5%
    447
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    665
    +0.10
    0.06
    217
    +0.10
    0.04
    126
    +0.10
    0.04
    Negative Logits
    assi
    -0.66
    rium
    -0.64
     Flavoring
    -0.62
    TON
    -0.61
    akia
    -0.59
    yip
    -0.59
     Rahman
    -0.58
     recruitment
    -0.56
     Fas
    -0.55
     Nguyen
    -0.54
    POSITIVE LOGITS
     differently
    1.19
     elsewhere
    1.03
     solely
    1.03
     concurrently
    1.03
     correctly
    1.01
     incorrectly
    0.99
     abroad
    0.97
     exclusively
    0.97
     indoors
    0.96
     cheaply
    0.95
    Activations Density 0.203%

    No Known Activations