© 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 768
    9. 8-RES_FS768-JB
    11. 749
    Prev
    Next
    INDEX
    Explanations

    expressions of gratitude and appreciation

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs768-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    768
    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
    467
    +0.13
    0.6%
    218
    +0.11
    0.5%
    542
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    105
    +0.13
    0.11
    467
    +0.11
    0.10
    521
    +0.10
    0.04
    Negative Logits
     predicate
    -0.76
     conventional
    -0.75
    ausible
    -0.74
     deflation
    -0.72
     plaus
    -0.72
     Worse
    -0.72
     narrower
    -0.71
     argues
    -0.70
     contends
    -0.67
     indist
    -0.67
    POSITIVE LOGITS
     THANK
    1.26
     congr
    1.20
     thank
    1.14
     Thank
    1.14
    Thank
    1.10
     gracious
    1.06
    thank
    1.01
     generosity
    1.00
     Congratulations
    0.99
    Congratulations
    0.99
    Activations Density 7.237%

    No Known Activations