© 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 24576
    9. 8-RES_FS24576-JB
    11. 19802
    Prev
    Next
    INDEX
    Explanations

    negative descriptions and critiques

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs24576-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    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
    314
    +0.17
    0.8%
    639
    +0.13
    0.6%
    756
    +0.13
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    314
    +0.17
    0.02
    152
    +0.13
    0.05
    639
    +0.13
    0.04
    Negative Logits
     flirt
    -0.69
     Dickinson
    -0.66
     Loud
    -0.62
     hiber
    -0.61
     stealth
    -0.60
     Codec
    -0.60
     caution
    -0.60
     Arabian
    -0.59
     Airlines
    -0.59
     troop
    -0.59
    POSITIVE LOGITS
    turned
    1.30
    cum
    1.21
    style
    1.17
    sized
    1.14
    based
    1.12
    type
    1.10
    selves
    1.09
    covered
    1.09
    inspired
    1.09
    sama
    1.08
    Activations Density 0.115%

    No Known Activations