Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    © Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 12288
    5. 8-RES_FS12288-JB
    6. 1043
    Prev
    Next
    INDEX
    Explanations

    Japanese words and phrases

    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
    116
    +0.13
    0.6%
    76
    +0.11
    0.5%
    183
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.13
    0.03
    738
    +0.11
    0.02
    76
    +0.09
    0.02
    Negative Logits
    quartered
    -0.86
    riott
    -0.66
     contrasting
    -0.64
     hitch
    -0.63
     horns
    -0.61
     strap
    -0.60
    bably
    -0.59
     conspicuous
    -0.59
     Pradesh
    -0.59
    ridges
    -0.58
    POSITIVE LOGITS
    arro
    1.44
    ombie
    1.14
    enegger
    1.07
    hou
    1.06
    hao
    1.01
    umi
    0.94
    akura
    0.89
    ymes
    0.86
    abal
    0.86
    awa
    0.86
    Activations Density 0.070%

    No Known Activations