© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Open Source Sparse Autoencoders for all Residual Stream Layers of GPT2-Small
    5. GPT2-Small
    7. Residual Stream
    9. 2-RES-JB
    11. 19457
    Prev
    Next
    INDEX
    Explanations

    There seems to be a variety of keywords, but the neuron is primarily activated by phrases that indicate an impressive or large amount, such as "whopping" or "staggering," irrespective of the context

    oai_token-act-pair · gpt-3.5-turbo

    phrases that emphasize large quantities or significant impact

    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
     Dise
    -0.84
    relations
    -0.80
     livest
    -0.79
    gemony
    -0.75
    ĪĴ
    -0.74
     Purg
    -0.73
    yer
    -0.72
     Trials
    -0.71
    NetMessage
    -0.70
    ©¶æ
    -0.70
    POSITIVE LOGITS
     whopping
    0.88
     amounts
    0.83
     amount
    0.80
    ulously
    0.80
     array
    0.70
     horsepower
    0.69
     totals
    0.69
     proportions
    0.68
     staggering
    0.67
     haul
    0.67
    Activations Density 0.014%

    No Known Activations