© 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 3072
    9. 8-RES_FS3072-JB
    11. 2965
    Prev
    Next
    INDEX
    Explanations

    website references

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    659
    +0.14
    0.7%
    332
    +0.12
    0.6%
    36
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.14
    0.11
    332
    +0.12
    0.09
    36
    +0.10
    0.07
    Negative Logits
     unrestricted
    -0.70
     CPR
    -0.67
     fret
    -0.64
     techno
    -0.63
     buzz
    -0.62
    tainment
    -0.61
     stressed
    -0.61
     DIY
    -0.60
     promotion
    -0.60
     subcontract
    -0.59
    POSITIVE LOGITS
    ixty
    1.44
    ESSION
    1.33
    ustain
    1.32
    ometime
    1.27
    esame
    1.25
    ourced
    1.25
    atisf
    1.23
    ixties
    1.22
    iblings
    1.22
    olutions
    1.22
    Activations Density 0.438%

    No Known Activations