© 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
    2. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    3. GPT2-Small
    4. Residual Post - 32k
    5. 8-RES_POST_32K-OAI
    6. 8258
    Prev
    Next
    INDEX
    Explanations

    variations of the word "ab" as well as related phonetic constructions

    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
    659
    +0.11
    0.5%
    10
    +0.10
    0.5%
    145
    +0.10
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    659
    +0.11
    0.07
    232
    +0.10
    0.07
    373
    +0.10
    -0.02
    Negative Logits
    UCT
    -0.65
    ãĥ»
    -0.65
     inaction
    -0.63
    ebted
    -0.61
    ãĥ¼ãĥĨãĤ£
    -0.58
    OWER
    -0.58
    ULAR
    -0.58
    FORM
    -0.57
    clips
    -0.55
     wrap
    -0.54
    POSITIVE LOGITS
    arov
    0.82
    ©¶æ
    0.81
    dain
    0.79
    udos
    0.78
    rys
    0.77
    uo
    0.77
    mine
    0.77
    atar
    0.76
    atsuki
    0.76
    abus
    0.75
    Activations Density 0.354%

    No Known Activations