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

    Neuronpedia

    APISteerSAE EvalsBlog/PodcastNEWSlackPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlog/PodcastGitHubSlackTwitterContact
    1. Home
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 2-RES-SM
    6. 29290
    Prev
    Next
    INDEX
    Explanations

    the word "But" indicating a contrast or contradiction

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__res-sm_processed/2-res-sm
    Prompts (Dashboard)
    32,768 prompts, 128 tokens each
    Dataset (Dashboard)
    monology/pile-uncopyrighted
    Features
    32,768
    Data Type
    torch.float32
    Hook Name
    blocks.2.hook_resid_post
    Hook Layer
    2
    Architecture
    standard
    Context Size
    128
    Dataset
    EleutherAI/the_pile_deduplicated
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    96
    +0.14
    0.8%
    369
    +0.13
    0.7%
    274
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    460
    +0.14
    0.05
    274
    +0.13
    0.05
    243
    +0.13
    0.05
    Negative Logits
     himself
    -1.93
    ersion
    -1.77
     herself
    -1.68
    )\]
    -1.68
    acs
    -1.63
    )\].
    -1.55
    ---|---|---
    -1.51
    ables
    -1.51
    etheus
    -1.49
    otin
    -1.47
    POSITIVE LOGITS
    tons
    2.88
    tig
    2.62
    thead
    1.97
    chers
    1.75
     hey
    1.63
     arresting
    1.61
    Į
    1.59
     jeopardy
    1.55
    cher
    1.55
     lately
    1.50
    Activations Density 0.229%

    No Known Activations