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. 29471
    Prev
    Next
    INDEX
    Explanations

    emotional emphasis within the text

    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
    396
    +0.13
    0.7%
    391
    +0.11
    0.6%
    381
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    18
    +0.13
    0.03
    391
    +0.11
    0.02
    446
    +0.11
    0.02
    Negative Logits
    rapped
    -1.53
    gebra
    -1.51
    ways
    -1.46
     homosexual
    -1.41
    googleapis
    -1.41
     fork
    -1.40
    atural
    -1.36
    ctions
    -1.35
    wegian
    -1.34
    fts
    -1.33
    POSITIVE LOGITS
    ulation
    1.70
    ulator
    1.68
    bourg
    1.67
    bers
    1.66
    ©
    1.63
    ulators
    1.63
    ulsions
    1.60
    ħ
    1.59
    ersion
    1.59
    erton
    1.57
    Activations Density 0.132%

    No Known Activations