© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    5. Pythia-70M-Deduped
    7. Residuals
    9. 2-RES-SM
    11. 29886
    Prev
    Next
    INDEX
    Explanations

    references to software licensing terms

    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
    157
    +0.13
    0.8%
    205
    +0.11
    0.6%
    2
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    329
    +0.13
    0.02
    157
    +0.11
    0.02
    205
    +0.11
    0.02
    Negative Logits
    Ł
    -1.98
    Īĺ
    -1.82
    ĵ
    -1.60
    ľĵ
    -1.60
    ĺ
    -1.53
    than
    -1.52
    Ĺ
    -1.50
    ĨĴ
    -1.49
    ¯
    -1.45
     breast
    -1.43
    POSITIVE LOGITS
     Agreement
    1.77
     agreement
    1.54
     waiver
    1.52
     expire
    1.50
     Certificate
    1.50
    âĦ
    1.47
    ом
    1.45
     identifiers
    1.45
    ÑĮ
    1.44
    framework
    1.41
    Activations Density 0.030%

    No Known Activations