© 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. 10995
    Prev
    Next
    INDEX
    Explanations

    terms related to various groups of people or audiences, such as travelers, fans, and visitors

    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
    156
    +0.25
    1.5%
    365
    +0.16
    1.0%
    32
    +0.14
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    365
    +0.25
    0.11
    32
    +0.16
    0.09
    64
    +0.14
    0.09
    Negative Logits
    ĥ½
    -4.76
     ↵
    -4.65
    ↵
    -4.65
    ↵
    -4.65
    -4.65
    ↵
    -4.65
    ↵
    -4.65
    -4.65
    -4.65
    -4.65
    POSITIVE LOGITS
    hip
    2.87
    pace
    2.64
    ervice
    2.35
    hell
    2.18
    iscus
    2.06
    hoe
    2.06
    hips
    1.98
    creen
    1.97
    ystems
    1.95
    ource
    1.90
    Activations Density 1.166%

    No Known Activations