© 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 EvalsExportsAPI 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. 6254
    Prev
    Next
    INDEX
    Explanations

    instances of the word "Chronicle" and associated visual content 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
    236
    +0.13
    0.6%
    717
    +0.09
    0.4%
    739
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    194
    +0.13
    0.01
    587
    +0.09
    0.01
    373
    +0.09
    -0.01
    Negative Logits
    cially
    -0.81
     Jenner
    -0.74
    cé
    -0.71
    onom
    -0.71
    lly
    -0.69
    ously
    -0.68
     Dob
    -0.68
    cest
    -0.66
    tics
    -0.64
    kamp
    -0.64
    POSITIVE LOGITS
    adr
    0.82
    apers
    0.77
    URES
    0.70
    arters
    0.70
    entric
    0.67
    afe
    0.67
     Chronicle
    0.67
    entry
    0.66
    izoph
    0.63
    alde
    0.63
    Activations Density 0.034%

    No Known Activations