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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 2-RES-SM
    6. 11856
    Prev
    Next
    INDEX
    Explanations

    the word "that" in various contexts

    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
    434
    +0.14
    0.8%
    290
    +0.13
    0.7%
    402
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    402
    +0.14
    0.07
    401
    +0.13
    0.07
    434
    +0.12
    0.07
    Negative Logits
    bat
    -1.63
     himself
    -1.56
     forums
    -1.50
    itud
    -1.50
     presentations
    -1.49
     forum
    -1.46
    them
    -1.43
     drama
    -1.41
     speeches
    -1.37
    course
    -1.34
    POSITIVE LOGITS
     overex
    1.62
     stray
    1.52
     populate
    1.51
    idis
    1.50
    lı
    1.50
     happened
    1.48
     contain
    1.43
    ched
    1.43
    of
    1.42
     borrow
    1.41
    Activations Density 0.300%

    No Known Activations