© 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. 10244
    Prev
    Next
    INDEX
    Explanations

    phrases related to business operations and processes

    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
    225
    +0.16
    0.7%
    724
    +0.14
    0.6%
    509
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    225
    +0.16
    0.10
    724
    +0.14
    0.09
    288
    +0.09
    0.08
    Negative Logits
    ãĥīãĥ©
    -0.81
    NL
    -0.76
    ighth
    -0.68
    Mos
    -0.67
    toc
    -0.66
    ãĥĺ
    -0.66
    zel
    -0.66
    batch
    -0.65
    cot
    -0.65
    ãĤ±
    -0.64
    POSITIVE LOGITS
     continuum
    0.88
     conducive
    0.82
     optimization
    0.80
     dictated
    0.79
     favoring
    0.77
     mismatch
    0.76
     dictates
    0.73
     influencing
    0.73
     flexibility
    0.73
     optimized
    0.70
    Activations Density 0.488%

    No Known Activations