© 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. 37230
    Prev
    Next
    INDEX
    Explanations

    adjectives that describe emotions, states, and qualities

    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
    584
    +0.14
    0.6%
    543
    +0.10
    0.4%
    707
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    72
    +0.14
    0.03
    197
    +0.10
    0.02
    151
    +0.09
    0.03
    Negative Logits
    Ô
    -0.66
     readings
    -0.63
    zsche
    -0.62
     circum
    -0.61
     secondary
    -0.59
    OPS
    -0.59
    Sem
    -0.58
    ighth
    -0.57
     registrations
    -0.57
    çĽ
    -0.56
    POSITIVE LOGITS
    amus
    0.73
    !!!!
    0.72
     ¯
    0.72
    !!!!!!!!
    0.71
    ranged
    0.71
     athlet
    0.71
    idious
    0.70
     lately
    0.69
    humane
    0.69
    atically
    0.69
    Activations Density 0.061%

    No Known Activations