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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsBlogSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Feature Splitting for GPT2-Small
    3. GPT2-Small
    4. Residual Feature Splitting 3072
    5. 8-RES_FS3072-JB
    6. 2178
    Prev
    Next
    INDEX
    Explanations

    phrases related to physical danger and conflict

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs3072-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    3,072
    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
    650
    +0.13
    0.6%
    218
    +0.11
    0.5%
    737
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    650
    +0.13
    0.10
    55
    +0.11
    0.11
    102
    +0.11
    0.11
    Negative Logits
     ourselves
    -0.85
     Families
    -0.80
    vae
    -0.75
     introduces
    -0.73
    inally
    -0.71
     Instruction
    -0.70
     oneself
    -0.68
    iatus
    -0.68
    heny
    -0.68
    ivals
    -0.67
    POSITIVE LOGITS
     hers
    1.17
     theirs
    0.96
     yours
    0.89
     priceless
    0.88
     mine
    0.84
     betray
    0.77
     dece
    0.74
     stained
    0.73
     amput
    0.72
    netflix
    0.71
    Activations Density 2.175%

    No Known Activations