© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    Jacobian LensNEW
    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsinterp-engineNEWAPI 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. 4153
    Prev
    Next
    INDEX
    Explanations

    phrases related to conditional scenarios and consequences

    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
    440
    +0.10
    0.5%
    546
    +0.10
    0.4%
    113
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    720
    +0.10
    0.15
    393
    +0.10
    0.14
    546
    +0.10
    0.09
    Negative Logits
     lately
    -0.73
    opus
    -0.64
     dated
    -0.63
    iasco
    -0.60
    requently
    -0.59
     recently
    -0.58
    recent
    -0.55
    kept
    -0.55
    ascus
    -0.55
     contrad
    -0.55
    POSITIVE LOGITS
    morrow
    1.06
    will
    0.96
     SHALL
    0.96
     will
    0.90
    Will
    0.84
     someday
    0.82
     be
    0.74
     WILL
    0.72
     Will
    0.72
     shall
    0.70
    Activations Density 2.244%

    No Known Activations