© 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 24576
    9. 8-RES_FS24576-JB
    11. 7171
    Prev
    Next
    INDEX
    Explanations

    words related to overall impressions or evaluations

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    513
    +0.11
    0.5%
    479
    +0.10
    0.4%
    260
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    513
    +0.11
    0.02
    299
    +0.10
    0.02
    623
    +0.10
    0.02
    Negative Logits
     Caf
    -0.77
     Debor
    -0.73
    lez
    -0.71
     Bry
    -0.69
    ÅŁ
    -0.68
    alde
    -0.67
    ©¶æ
    -0.66
    lict
    -0.65
    chers
    -0.65
     Claud
    -0.65
    POSITIVE LOGITS
     usefulness
    0.91
     stability
    0.89
     attractiveness
    0.88
     momentum
    0.88
     effectiveness
    0.87
     length
    0.84
     improvement
    0.82
     strategy
    0.81
     impression
    0.80
     efficiency
    0.79
    Activations Density 0.025%

    No Known Activations