© 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. 26540
    Prev
    Next
    INDEX
    Explanations

    phrases and words related to the impact of actions on a collective or societal level

    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
    43
    +0.12
    0.5%
    373
    +0.10
    0.5%
    679
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    43
    +0.12
    0.03
    596
    +0.10
    0.02
    384
    +0.10
    0.02
    Negative Logits
    minster
    -0.73
    thal
    -0.66
    ishly
    -0.65
    ãĥĥãĥī
    -0.64
    corn
    -0.63
    iously
    -0.62
    eps
    -0.60
    ragon
    -0.60
    phony
    -0.60
    glass
    -0.59
    POSITIVE LOGITS
     however
    1.01
     though
    0.82
     insofar
    0.77
     we
    0.76
     let
    0.76
     there
    0.75
     moreover
    0.74
     they
    0.72
     lest
    0.71
     remember
    0.69
    Activations Density 0.053%

    No Known Activations