© 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. 37283
    Prev
    Next
    INDEX
    Explanations

    phrases related to negative consequences or costs incurred

    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
    247
    +0.12
    0.6%
    285
    +0.11
    0.5%
    344
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    539
    +0.12
    0.01
    247
    +0.11
    0.01
    285
    +0.11
    0.01
    Negative Logits
    abad
    -0.75
    ortium
    -0.71
    kan
    -0.71
    oub
    -0.69
    oggles
    -0.67
    packed
    -0.66
     Clover
    -0.65
    rete
    -0.65
    alde
    -0.65
    binding
    -0.64
    POSITIVE LOGITS
     detriment
    0.90
     expense
    0.78
     altar
    0.76
     thereof
    0.76
    othal
    0.70
    liest
    0.69
     inconvenient
    0.65
    ħĭ
    0.64
     disadvantage
    0.63
     taxpayers
    0.63
    Activations Density 0.013%

    No Known Activations