© 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. 41696
    Prev
    Next
    INDEX
    Explanations

    references to third-party entities or organizations

    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
    152
    +0.13
    0.6%
    460
    +0.12
    0.5%
    506
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    152
    +0.13
    0.03
    460
    +0.12
    0.03
    425
    +0.10
    0.03
    Negative Logits
    Ń·
    -0.78
    pload
    -0.74
     Canaver
    -0.68
    ++++++++++++++++
    -0.67
     Richards
    -0.64
     Restrict
    -0.63
     Manson
    -0.62
     Maxim
    -0.61
     Limits
    -0.60
     confines
    -0.60
    POSITIVE LOGITS
    party
    1.28
    degree
    1.20
    tier
    1.14
    dimensional
    1.12
    generation
    1.05
    level
    1.00
    century
    0.99
    rate
    0.96
    country
    0.94
    chance
    0.93
    Activations Density 0.016%

    No Known Activations