© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 49152
    9. 8-RES_FS49152-JB
    11. 42928
    Prev
    Next
    INDEX
    Explanations

    instances of the indefinite article "a" and its frequency in different contexts

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs49152-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    49,152
    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
    36
    +0.21
    1.0%
    393
    +0.14
    0.7%
    263
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    36
    +0.21
    0.04
    636
    +0.14
    0.02
    263
    +0.10
    0.02
    Negative Logits
    )=(
    -0.81
    alian
    -0.76
    achu
    -0.76
    pard
    -0.75
    çī
    -0.72
    anism
    -0.72
    aly
    -0.70
    adj
    -0.70
    Enjoy
    -0.70
    çĶ°
    -0.69
    POSITIVE LOGITS
     doubt
    1.05
     clue
    1.01
     lot
    0.98
     hitch
    0.98
     hint
    0.97
     definite
    0.92
     proper
    0.91
     grasp
    0.91
     guarantee
    0.90
     corresponding
    0.88
    Activations Density 0.026%

    No Known Activations