© 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. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    5. GPT2-Small
    7. Residual Post - 32k
    9. 8-RES_POST_32K-OAI
    11. 5449
    Prev
    Next
    INDEX
    Explanations

    expressions of gratitude and feedback in communication

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_8.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.8.hook_resid_post
    Hook Layer
    8
    Architecture
    standard
    Context Size
    64
    Dataset
    Skylion007/openwebtext
    Activation Function
    topk
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    71
    +0.13
    0.6%
    756
    +0.13
    0.6%
    752
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    86
    +0.13
    0.12
    705
    +0.13
    0.15
    756
    +0.10
    -0.15
    Negative Logits
     nerds
    -0.81
     gadgets
    -0.79
     grinning
    -0.76
    toggle
    -0.75
     shrug
    -0.75
     booze
    -0.74
     underdog
    -0.73
     dudes
    -0.72
     reinvent
    -0.72
     perk
    -0.71
    POSITIVE LOGITS
    sic
    1.10
     Defendant
    0.97
    Senator
    0.97
     Plaintiff
    0.97
     regrett
    0.92
     Defendants
    0.91
     Counsel
    0.91
    approximately
    0.88
     pursuant
    0.87
    Statement
    0.87
    Activations Density 14.760%

    No Known Activations