© 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. 12943
    Prev
    Next
    INDEX
    Explanations

    instances of the word "would" indicating potential or hypothetical actions

    oai_token-act-pair · gpt-3.5-turbo

    instances of the word "would" indicating intention or hypothetical situations

    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
    rylic
    -0.70
    æ³
    -0.66
    holder
    -0.64
    values
    -0.64
    eyed
    -0.64
    Reporting
    -0.62
    Ti
    -0.62
    isted
    -0.61
     Kag
    -0.60
    meta
    -0.60
    POSITIVE LOGITS
     be
    1.13
     surely
    1.03
     doubtless
    1.01
     never
    1.00
     inevitably
    0.98
     undoubtedly
    0.96
     eventually
    0.95
    ĸļ
    0.92
     suffice
    0.92
     probably
    0.89
    Activations Density 0.151%

    No Known Activations