© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it
    Neuronpedia
    Jacobian LensNEW
    Natural Language
    Autoencoders
    MAY 26    Assistant AxisJAN 26Circuit TracerAPR 26SteerSAE EvalsExportsinterp-engineNEW 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. 26381
    Prev
    Next
    INDEX
    Explanations

    prepositions and their associated phrases in relation to spatial positions

    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
    98
    +0.20
    1.1%
    156
    +0.14
    0.8%
    10
    +0.13
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    10
    +0.20
    0.16
    291
    +0.14
    0.18
    98
    +0.13
    0.04
    Negative Logits
    atives
    -1.51
     certificates
    -1.42
    strom
    -1.41
    ially
    -1.34
    ableness
    -1.34
    ative
    -1.34
    ional
    -1.34
    wise
    -1.32
    ive
    -1.32
    ↵
    -1.31
    POSITIVE LOGITS
    »¿
    2.83
    ĨĴ
    2.32
    ľĵ
    1.98
    Ľ
    1.93
    ĻĤ
    1.91
    Ĥ¬
    1.91
    ¿
    1.90
    IJ
    1.89
    rapy
    1.84
    ĸ
    1.80
    Activations Density 1.456%

    No Known Activations