© 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. 28311
    Prev
    Next
    INDEX
    Explanations

    terms related to evaluating and measuring behavior or activities

    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
    186
    +0.19
    1.1%
    240
    +0.16
    0.9%
    5
    +0.15
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    204
    +0.19
    0.14
    292
    +0.16
    0.13
    256
    +0.15
    0.14
    Negative Logits
     basis
    -1.90
    ware
    -1.70
    aurus
    -1.52
     jurisdiction
    -1.46
     shores
    -1.45
    atur
    -1.41
    bilt
    -1.40
     citations
    -1.40
    atura
    -1.36
     avenues
    -1.34
    POSITIVE LOGITS
    Ī
    2.21
    Ł
    2.00
    ĸ
    1.96
    ĻĤ
    1.93
    ¥
    1.93
    ł
    1.91
    Ģ
    1.90
    ŀ
    1.88
    ¼
    1.82
    ľ
    1.82
    Activations Density 5.186%

    No Known Activations