© 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. 9788
    Prev
    Next
    INDEX
    Explanations

    phrases indicating movement or passage through a space

    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
    412
    +0.15
    0.9%
    121
    +0.14
    0.8%
    45
    +0.12
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    412
    +0.15
    0.08
    121
    +0.14
    0.06
    16
    +0.12
    0.06
    Negative Logits
    §
    -4.08
    «
    -3.70
    Ļ
    -3.40
    ®
    -3.39
    ŀ
    -3.38
    ī
    -3.36
    ĻĤ
    -3.32
    Ī
    -3.32
    ²
    -3.30
    ¥
    -3.30
    POSITIVE LOGITS
     injured
    1.80
    ventory
    1.60
     disguise
    1.53
    dden
    1.53
    robe
    1.51
    htra
    1.51
    ager
    1.49
    board
    1.47
    ockets
    1.47
     corners
    1.47
    Activations Density 0.800%

    No Known Activations