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

    Neuronpedia

    APIAssistant AxisNEWCircuit 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. 29184
    Prev
    Next
    INDEX
    Explanations

    terms related to systematic approaches

    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
    198
    +0.15
    0.8%
    92
    +0.11
    0.6%
    327
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    198
    +0.15
    0.02
    148
    +0.11
    0.00
    424
    +0.11
    0.01
    Negative Logits
     mismo
    -1.93
    aceae
    -1.70
    inois
    -1.69
     skilled
    -1.66
     behalf
    -1.52
     pills
    -1.46
     brave
    -1.42
    inea
    -1.41
    owed
    -1.40
    wegian
    -1.40
    POSITIVE LOGITS
    ignment
    1.72
    ity
    1.57
    °
    1.57
    "}](#
    1.50
     underest
    1.50
    ķ
    1.47
    essment
    1.46
    itat
    1.43
    ities
    1.42
    sheet
    1.42
    Activations Density 0.147%

    No Known Activations