© 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 EvalsExportsinterp-engineNEWAPI 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. 13972
    Prev
    Next
    INDEX
    Explanations

    specific acronyms, names, or entities associated with certain groups or organizations

    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
    17
    +0.25
    1.5%
    231
    +0.13
    0.8%
    186
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    17
    +0.25
    0.02
    82
    +0.13
    0.05
    231
    +0.12
    0.07
    Negative Logits
    lessly
    -1.76
    msgstr
    -1.49
    aily
    -1.47
     deaf
    -1.38
    ographers
    -1.37
    iffs
    -1.33
    atter
    -1.33
    ims
    -1.33
    forth
    -1.31
    arers
    -1.30
    POSITIVE LOGITS
    1.85
    1.85
    1.85
    ↵
    1.85
    1.85
    <|padding|>
    1.85
    ↵Č
    1.85
    ↵            ↵
    1.85
    <|outofrange|>
    1.85
    1.85
    Activations Density 0.737%

    No Known Activations