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

    Neuronpedia

    Natural Language
    Autoencoders
    NEW    Assistant AxisNEWCircuit TracerUPDATESteerSAE EvalsExportsAPI 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. 29991
    Prev
    Next
    INDEX
    Explanations

    references to the United States and its legal cases

    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
    199
    +0.14
    0.7%
    410
    +0.12
    0.7%
    34
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    34
    +0.14
    0.02
    168
    +0.12
    0.01
    213
    +0.12
    0.01
    Negative Logits
    ...](
    -1.71
    ishing
    -1.53
     starvation
    -1.52
     starving
    -1.51
     untreated
    -1.48
     buckets
    -1.46
    tons
    -1.46
    iph
    -1.43
     generations
    -1.43
     generation
    -1.42
    POSITIVE LOGITS
    ²
    3.76
    »¿
    3.66
    ģ
    3.40
    Īĺ
    3.39
    °
    3.34
    Ī
    3.27
    ¸
    3.23
    Ļª
    3.20
    ĸ´
    3.18
    ·
    3.12
    Activations Density 0.023%

    No Known Activations