© 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. 4-RES-JB
    11. 9170
    Prev
    Next
    INDEX
    Explanations

    phrases indicating preference or speculative statements

    oai_token-act-pair · gpt-3.5-turbo

    the conditional phrase "would" indicating hypothetical or future actions

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 4-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    4
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
     Kag
    -0.67
    building
    -0.62
    values
    -0.60
     Gutenberg
    -0.60
    è¦ļéĨĴ
    -0.60
    Reporting
    -0.59
    Ready
    -0.59
    resources
    -0.59
    DNA
    -0.59
    çİĭ
    -0.58
    POSITIVE LOGITS
     surely
    0.94
     undoubtedly
    0.90
     be
    0.90
     doubtless
    0.89
     gladly
    0.89
     definitely
    0.86
     inevitably
    0.86
     suffice
    0.83
     never
    0.83
     probably
    0.83
    Activations Density 0.155%

    No Known Activations