© 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. 29852
    Prev
    Next
    INDEX
    Explanations

    expressions related to mathematical operations and calculations

    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
    221
    +0.14
    0.7%
    74
    +0.12
    0.7%
    79
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    200
    +0.14
    0.02
    221
    +0.12
    0.02
    79
    +0.12
    0.02
    Negative Logits
     haven
    -1.78
    acea
    -1.65
    iph
    -1.53
    /)
    -1.53
    /'
    -1.50
     aren
    -1.48
    .[]{
    -1.48
     intended
    -1.47
     don
    -1.46
    una
    -1.45
    POSITIVE LOGITS
    ¦
    2.07
    ©
    1.79
     Exercise
    1.54
    ĻĤ
    1.54
    ķ
    1.53
    helm
    1.49
    ħ
    1.48
     degrees
    1.46
    ŀ
    1.44
     Kaiser
    1.44
    Activations Density 0.113%

    No Known Activations