© 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 49152
    9. 8-RES_FS49152-JB
    11. 41738
    Prev
    Next
    INDEX
    Explanations

    the occurrence of the letters "th" in various 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
    496
    +0.10
    0.5%
    678
    +0.10
    0.4%
    701
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    701
    +0.10
    0.01
    742
    +0.10
    0.01
    280
    +0.09
    0.01
    Negative Logits
    GMT
    -0.77
    ITED
    -0.66
    ships
    -0.59
    UTION
    -0.57
     Clash
    -0.56
     PTSD
    -0.56
    ++++++++++++++++
    -0.56
     Porter
    -0.55
     Chase
    -0.55
     Wonderland
    -0.55
    POSITIVE LOGITS
    ousands
    1.33
    umbnail
    1.30
    irteen
    1.23
    ousand
    1.22
    irty
    1.17
    ulhu
    1.13
    ttp
    1.08
    ority
    1.07
    ongs
    1.06
    ieving
    1.06
    Activations Density 0.008%

    No Known Activations