© 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. 10170
    Prev
    Next
    INDEX
    Explanations

    mention of failures or negative outcomes within a context

    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
    441
    +0.10
    0.4%
    472
    +0.09
    0.4%
    274
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    188
    +0.10
    0.02
    472
    +0.09
    0.02
    78
    +0.09
    0.02
    Negative Logits
    bern
    -0.80
    utra
    -0.77
    selves
    -0.71
    riel
    -0.71
    ieri
    -0.67
    wick
    -0.66
    upt
    -0.65
    ordon
    -0.65
    Edge
    -0.65
    rete
    -0.64
    POSITIVE LOGITS
     miser
    1.10
     failures
    0.81
     rate
    0.80
    lust
    0.77
    Failure
    0.75
     derail
    0.72
     failure
    0.70
     outcome
    0.69
     dism
    0.68
     occurrence
    0.68
    Activations Density 0.066%

    No Known Activations