© 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. 343
    Prev
    Next
    INDEX
    Explanations

    praise or critique regarding the performance or qualities of individuals

    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
    354
    +0.13
    0.6%
    139
    +0.13
    0.6%
    762
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    139
    +0.13
    0.11
    354
    +0.13
    0.09
    762
    +0.11
    0.09
    Negative Logits
     Additional
    -0.66
    CCC
    -0.59
     affiliate
    -0.59
     Joint
    -0.59
    grave
    -0.58
    entry
    -0.58
    earch
    -0.56
    prem
    -0.55
    1007
    -0.55
    SE
    -0.55
    POSITIVE LOGITS
    zbollah
    1.23
    'll
    1.14
    pherd
    1.11
     certainly
    1.07
     ain
    1.06
    gemony
    1.04
     didn
    1.03
    ldon
    1.01
     loves
    0.98
     didnt
    0.98
    Activations Density 0.377%

    No Known Activations