© 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. 31521
    Prev
    Next
    INDEX
    Explanations

    the preposition "at"

    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
    369
    +0.13
    0.7%
    222
    +0.11
    0.6%
    34
    +0.10
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    222
    +0.13
    0.01
    420
    +0.11
    0.01
    181
    +0.10
    0.01
    Negative Logits
    ands
    -1.56
    iratory
    -1.47
    apers
    -1.42
    vents
    -1.40
    assisted
    -1.39
    ?’
    -1.38
    insula
    -1.38
    aptic
    -1.38
    ANE
    -1.37
    osse
    -1.36
    POSITIVE LOGITS
     unpublished
    1.76
    mosp
    1.62
     least
    1.55
     multiplicity
    1.44
    Lic
    1.44
     award
    1.42
     cod
    1.41
     rpm
    1.39
     license
    1.39
    wright
    1.36
    Activations Density 0.041%

    No Known Activations