© 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. 4780
    Prev
    Next
    INDEX
    Explanations

    patterns related to technical commands and file management

    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
    98
    +0.12
    0.5%
    296
    +0.11
    0.5%
    262
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    326
    +0.12
    0.08
    696
    +0.11
    0.04
    396
    +0.10
    0.02
    Negative Logits
     horm
    -0.74
     coil
    -0.72
    uld
    -0.70
    ctors
    -0.70
     masc
    -0.69
     fend
    -0.69
     creature
    -0.67
     handic
    -0.67
     chirop
    -0.67
    ollah
    -0.66
    POSITIVE LOGITS
     Online
    1.06
     Updates
    1.02
    Update
    0.98
    ~
    0.97
     ~
    0.96
    Updated
    0.94
    Initialized
    0.94
     Available
    0.93
    update
    0.93
     updated
    0.93
    Activations Density 1.060%

    No Known Activations