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. 4237
    Prev
    Next
    INDEX
    Explanations

    the term "solid" and its variations, indicating a focus on strength or quality

    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
    642
    +0.15
    0.7%
    622
    +0.10
    0.5%
    748
    +0.09
    0.4%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    622
    +0.15
    0.02
    93
    +0.10
    0.01
    207
    +0.09
    0.01
    Negative Logits
    Pause
    -0.76
     Kafka
    -0.73
     Tears
    -0.72
     Pigs
    -0.71
     Outbreak
    -0.69
     Zoo
    -0.69
    zie
    -0.68
     Mania
    -0.66
     Lovely
    -0.65
     Dy
    -0.65
    POSITIVE LOGITS
    ifying
    0.93
    ified
    0.91
    enough
    0.90
     immersion
    0.86
    ties
    0.81
    ior
    0.81
     groundwork
    0.79
    eenth
    0.78
    lly
    0.78
    ded
    0.77
    Activations Density 0.045%

    No Known Activations