© 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. 4965
    Prev
    Next
    INDEX
    Explanations

    possessive pronouns and their references

    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
    555
    +0.10
    0.5%
    117
    +0.09
    0.4%
    151
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    555
    +0.10
    0.04
    362
    +0.09
    0.04
    15
    +0.09
    0.04
    Negative Logits
    LOCK
    -0.79
    mination
    -0.79
    TABLE
    -0.74
    trop
    -0.70
    代
    -0.69
    mut
    -0.68
    LIN
    -0.67
    UL
    -0.66
    DH
    -0.66
    éĹ
    -0.66
    POSITIVE LOGITS
     il
    1.32
     allies
    1.15
     cohorts
    1.07
     associates
    1.04
     pals
    0.99
     affiliates
    0.94
     collaborators
    0.92
     colleagues
    0.91
     partners
    0.89
     companions
    0.89
    Activations Density 0.166%

    No Known Activations