Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. OpenAI · Sparse Autoencoder for GPT2-Small - v5 2024
    3. GPT2-Small
    4. Residual Post - 32k
    5. 6-RES_POST_32K-OAI
    6. 4215
    Prev
    Next
    INDEX
    Explanations

    the word "must" and its variations that indicate necessity or obligation

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 6-res_post_32k-oai
    Configuration
    jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_6.pt
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.6.hook_resid_post
    Hook Layer
    6
    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
    373
    +0.13
    0.6%
    306
    +0.13
    0.6%
    176
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    551
    +0.13
    0.02
    306
    +0.13
    0.02
    386
    +0.12
    0.02
    Negative Logits
    EStreamFrame
    -0.84
    é¾
    -0.68
    issance
    -0.67
    äºĶ
    -0.65
    ļéĨĴ
    -0.64
    ãģ®å®
    -0.64
    é»Ĵ
    -0.63
    uyomi
    -0.62
     tether
    -0.61
    ilater
    -0.61
    POSITIVE LOGITS
    reads
    0.92
    oms
    0.86
    airs
    0.82
    afi
    0.79
    inent
    0.78
    hew
    0.77
    fields
    0.76
    ens
    0.76
    croft
    0.76
    ocratic
    0.74
    Activations Density 0.032%

    No Known Activations