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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Apollo Research · Taylor · Identifying Functionally Important Features with End-to-End Sparse Dictionary Learning
    3. GPT2-Small
    4. Residual Similar L0 E2E Recon
    5. 2-RES_SLEFR-AJT
    6. 9863
    Prev
    Next
    INDEX
    Explanations

    The text shows various topics, often featuring quotes and providing commentary on events, debates, and opinions, highlighting contrasting perspectives and societal issues.

    eleuther_acts_top20 · gpt-4o-miniTriggered by @johnny

    details related to political and social issues.

    oai_token-act-pair · gpt-3.5-turboTriggered by @johnny

    A newline character, typically used to separate paragraphs or sections of text.

    eleuther_acts_top20 · gemini-1.5-proTriggered by @johnny

    references to specific individuals in a contextual narrative

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Configuration
    neuronpedia/gpt2-small__res_slefr-ajt/2-res_slefr-ajt
    Prompts (Dashboard)
    12,288 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    46,080
    Data Type
    torch.float32
    Hook Point
    blocks.2.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    apollo-research/Skylion007-openwebtext-tokenizer-gpt2
    Hook Point Layer
    2
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    640
    +0.36
    1.8%
    266
    +0.28
    1.3%
    160
    +0.16
    0.8%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    640
    +0.36
    0.53
    160
    +0.28
    0.38
    441
    +0.16
    0.23
    Negative Logits
    apon
    -0.76
    otine
    -0.71
    hei
    -0.70
    ardless
    -0.68
    utic
    -0.67
    icer
    -0.67
     girlfriends
    -0.67
    uly
    -0.65
     Saiyan
    -0.65
    itory
    -0.64
    POSITIVE LOGITS
    ³³³
    0.98
    Newsletter
    0.98
    Anyway
    0.94
    ³³³³
    0.92
    SPONSORED
    0.91
    Likewise
    0.91
    Conclusion
    0.91
    However
    0.90
    Meanwhile
    0.90
    Furthermore
    0.89
    Activations Density 9.043%

    No Known Activations