© 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. MLP Post
    9. 2-MLP-SM
    11. 25274
    Prev
    Next
    INDEX
    Explanations

    conjunctions and logical connectors in sentences

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    ctigges/pythia-70m-deduped__mlp-sm_processed/2-mlp-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_mlp_out
    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
    218
    +0.14
    0.8%
    253
    +0.14
    0.8%
    330
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    218
    +0.14
    0.06
    97
    +0.14
    0.05
    261
    +0.12
    0.03
    Negative Logits
    úblic
    -1.55
    chte
    -1.47
    atra
    -1.47
    scribe
    -1.37
    zip
    -1.37
    edia
    -1.35
    posure
    -1.35
    azine
    -1.33
    liv
    -1.30
    izoph
    -1.30
    POSITIVE LOGITS
     Exhibit
    1.63
    irs
    1.47
     denominator
    1.46
    uary
    1.44
    hold
    1.38
    ICAN
    1.32
     Deal
    1.32
    ++
    1.31
    GER
    1.31
    holds
    1.31
    Activations Density 0.328%

    No Known Activations