© 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. 13082
    Prev
    Next
    INDEX
    Explanations

    themes related to art and its intersection with video games as a form of learning and teaching

    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
    271
    +0.20
    1.2%
    479
    +0.16
    1.0%
    261
    +0.15
    0.9%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    261
    +0.20
    0.07
    479
    +0.16
    0.07
    140
    +0.15
    0.10
    Negative Logits
    ctomy
    -1.68
     remedy
    -1.45
    atism
    -1.44
     argument
    -1.40
     survives
    -1.40
     anomaly
    -1.38
     \%
    -1.36
     tumour
    -1.36
    ctrine
    -1.36
    AML
    -1.35
    POSITIVE LOGITS
    ĵ
    2.44
    Ĭ
    2.44
    Ŀ
    2.38
    Ĵ
    2.37
    ŀ
    2.36
    £
    2.30
    ¼
    2.29
    ĭ
    2.28
    Īĺ
    2.28
    ³
    2.27
    Activations Density 5.835%

    No Known Activations