Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
    3. Pythia-70M-Deduped
    4. Residuals
    5. 2-RES-SM
    6. 29175
    Prev
    Next
    INDEX
    Explanations

    greetings and welcoming phrases

    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
    156
    +0.22
    1.3%
    263
    +0.15
    0.9%
    302
    +0.13
    0.7%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    156
    +0.22
    0.42
    111
    +0.15
    0.41
    71
    +0.13
    0.40
    Negative Logits
    ħ
    -1.74
    ¯
    -1.67
    ŀ
    -1.63
    Ļ
    -1.58
    ĥ
    -1.57
    ¾
    -1.53
    ĸ
    -1.52
    Ŀ
    -1.51
     &=
    -1.50
    ģ
    -1.49
    POSITIVE LOGITS
     chat
    1.82
     congrat
    1.80
     yours
    1.78
     welcome
    1.68
     joking
    1.66
     Welcome
    1.61
     my
    1.59
     questions
    1.57
     comments
    1.56
     hello
    1.55
    Activations Density 1.609%

    No Known Activations