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

    Neuronpedia

    Jacobian LensNEW
    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. 9954
    Prev
    Next
    INDEX
    Explanations

    patterns and relationships involving actions or events over time

    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
    501
    +0.14
    0.8%
    269
    +0.12
    0.7%
    174
    +0.11
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    501
    +0.14
    0.03
    269
    +0.12
    0.03
    484
    +0.11
    0.02
    Negative Logits
    inct
    -1.54
     compatibility
    -1.42
    ory
    -1.37
    ary
    -1.36
     converts
    -1.30
    izability
    -1.30
     CRC
    -1.28
    onium
    -1.27
    ľĵ
    -1.27
    "|
    -1.25
    POSITIVE LOGITS
     på
    2.38
     må
    2.26
    för
    2.24
     så
    2.19
     eller
    2.19
     få
    2.06
     år
    2.06
     hå
    2.06
    vik
    2.04
     frå
    2.01
    Activations Density 0.282%

    No Known Activations