© 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. 27019
    Prev
    Next
    INDEX
    Explanations

    technical and programming terms related to data structures and operations

    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
    +0.39
    2.3%
    240
    +0.13
    0.8%
    17
    +0.11
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    17
    +0.39
    0.01
    90
    +0.13
    0.04
    376
    +0.11
    0.01
    Negative Logits
    ↵
    -3.78
    -3.78
    ↵↵
    -3.78
    <|outofrange|>
    -3.78
    -3.78
    -3.78
    č↵
    -3.78
    -3.78
    <|outofrange|>
    -3.78
    <|outofrange|>
    -3.78
    POSITIVE LOGITS
    pace
    2.84
    ystems
    2.72
    hell
    2.62
    ource
    2.52
    heet
    2.52
    mith
    2.49
    ystem
    2.45
    cript
    2.41
    ymbol
    2.37
    erver
    2.24
    Activations Density 0.496%

    No Known Activations