© 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
    2. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    3. GPT2-Small
    4. Residual Post - 32k
    5. 8-RES_POST_32K-OAI
    6. 4950
    Prev
    Next
    INDEX
    Explanations

    references to "man" and "men" in various contexts

    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
    515
    +0.14
    0.6%
    291
    +0.10
    0.5%
    430
    +0.10
    0.5%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    515
    +0.14
    0.05
    187
    +0.10
    0.04
    264
    +0.10
    0.03
    Negative Logits
    IVERS
    -0.77
    PsyNetMessage
    -0.67
    Import
    -0.64
    VICE
    -0.61
    REC
    -0.60
    Integ
    -0.60
    ItemTracker
    -0.59
    TING
    -0.58
    BLIC
    -0.57
    irtual
    -0.57
    POSITIVE LOGITS
    hunt
    1.38
    hood
    1.30
    nered
    1.23
    uscript
    1.16
    gling
    1.15
    hattan
    1.07
    agers
    1.02
    ager
    1.01
    fred
    0.99
    volent
    0.97
    Activations Density 0.122%

    No Known Activations