© 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. 6829
    Prev
    Next
    INDEX
    Explanations

    the word "to" followed by a verb

    oai_token-act-pair · gpt-3.5-turbo

    phrases indicating agreement or consent

    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
    patch
    -0.71
     wastes
    -0.66
    grain
    -0.65
     Detected
    -0.65
     Wonders
    -0.64
    clips
    -0.64
    ults
    -0.63
    wig
    -0.62
    blocks
    -0.60
     weeds
    -0.59
    POSITIVE LOGITS
     waive
    0.99
     cooperate
    0.98
     obey
    0.97
     accept
    0.96
     abide
    0.90
     participate
    0.89
     rehe
    0.87
     abandon
    0.83
     receive
    0.82
     collaborate
    0.81
    Activations Density 0.094%

    No Known Activations