© 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. 5442
    Prev
    Next
    INDEX
    Explanations

    phrases and words related to topics of discussion and biographical content

    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
    362
    +0.10
    0.4%
    414
    +0.09
    0.4%
    641
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    362
    +0.10
    0.10
    669
    +0.09
    0.07
    641
    +0.09
    0.07
    Negative Logits
    beam
    -0.74
     detector
    -0.69
     detectors
    -0.68
     briefed
    -0.66
    finger
    -0.66
    IER
    -0.65
     notification
    -0.64
     directive
    -0.64
    pour
    -0.64
    Request
    -0.64
    POSITIVE LOGITS
     mankind
    0.81
    aturday
    0.80
     humankind
    0.78
     humanity
    0.77
     colonialism
    0.75
     atheism
    0.75
     slavery
    0.74
     oppression
    0.73
     religion
    0.73
     wartime
    0.72
    Activations Density 1.032%

    No Known Activations