© 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. 4821
    Prev
    Next
    INDEX
    Explanations

    financial impacts of housing costs and economic policies on workers and employers

    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
    720
    +0.13
    0.6%
    350
    +0.11
    0.5%
    468
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    86
    +0.13
    0.13
    608
    +0.11
    0.21
    720
    +0.10
    0.18
    Negative Logits
    canon
    -1.14
    CVE
    -0.83
    RFC
    -0.83
     confir
    -0.83
     Fn
    -0.82
     PCR
    -0.82
     Detection
    -0.80
     Dates
    -0.79
     scrut
    -0.79
     SCP
    -0.77
    POSITIVE LOGITS
     incomes
    1.58
     millionaires
    1.47
     wealthier
    1.46
     poverty
    1.44
     unemployment
    1.43
     income
    1.38
     subsidized
    1.35
     unemployed
    1.35
     pensions
    1.35
     affluent
    1.34
    Activations Density 7.257%

    No Known Activations