Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 49152
    5. 8-RES_FS49152-JB
    6. 43210
    Prev
    Next
    INDEX
    Explanations

    gerunds and present participles, indicating ongoing actions or states

    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
    9
    +0.12
    0.5%
    195
    +0.11
    0.5%
    240
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    9
    +0.12
    0.01
    453
    +0.11
    0.01
    181
    +0.10
    0.01
    Negative Logits
    paio
    -0.76
    £ı
    -0.70
    çīĪ
    -0.68
     Jav
    -0.68
     cabinets
    -0.66
    ãĥīãĥ©
    -0.66
    INO
    -0.65
    tein
    -0.64
     gallery
    -0.64
    INC
    -0.64
    POSITIVE LOGITS
    etheless
    0.81
    itized
    0.76
    thood
    0.73
    bye
    0.73
    mers
    0.71
    theless
    0.71
    ication
    0.71
    efully
    0.70
    ibal
    0.70
    eful
    0.68
    Activations Density 0.010%

    No Known Activations