© 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
    2. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    3. GPT2-Small
    4. Residual Post - 32k
    5. 8-RES_POST_32K-OAI
    6. 4868
    Prev
    Next
    INDEX
    Explanations

    instances of negation or expressions of doubt

    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
    20
    +0.12
    0.5%
    373
    +0.11
    0.5%
    630
    +0.11
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    20
    +0.12
    0.11
    398
    +0.11
    0.09
    126
    +0.11
    0.07
    Negative Logits
     creations
    -0.65
    atic
    -0.61
     rotated
    -0.61
    achi
    -0.57
    arts
    -0.56
     lovers
    -0.54
    eps
    -0.54
     spor
    -0.54
     endeavors
    -0.52
    letes
    -0.52
    POSITIVE LOGITS
     ample
    0.91
    ibaba
    0.90
     room
    0.88
     unanim
    0.86
     plenty
    0.85
    enough
    0.81
     enough
    0.78
     disagreement
    0.76
    omething
    0.75
     overlap
    0.74
    Activations Density 0.424%

    No Known Activations