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

    Neuronpedia

    APICircuit TracerNEWSteerSAE EvalsExportsSlackBlogPrivacy & TermsContact
    © Neuronpedia 2025
    Privacy & TermsBlogGitHubSlackTwitterContact
    1. Home
    2. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    3. GPT2-Small
    4. Residual Stream
    5. 2-RES-JB
    6. 156
    Prev
    Next
    INDEX
    Explanations

    the adverb "badly" modifying another word

    oai_token-act-pair · gpt-4-turbo

    adverbs/adjective that describe the degree of negative impact or quality of something

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "badly" indicating negative situations or conditions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 2-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.hook_resid_pre
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    24,576
    Data Type
    torch.float32
    Hook Point
    blocks.2.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    2
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    anwhile
    -0.81
    heny
    -0.70
    eer
    -0.67
    orney
    -0.64
    utral
    -0.64
    ory
    -0.64
    ership
    -0.63
    ynthesis
    -0.63
    án
    -0.63
    scl
    -0.62
    POSITIVE LOGITS
     enough
    0.81
    WARE
    0.79
     behaved
    0.78
     badly
    0.74
    asses
    0.73
     damaged
    0.73
     patched
    0.72
    ãĤª
    0.70
     suited
    0.70
     decomp
    0.70
    Activations Density 0.006%

    No Known Activations