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

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & 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. 8-RES-JB
    6. 9175
    Prev
    Next
    INDEX
    Explanations

    the word "to" where it is expressed in a negative context or as a fault

    oai_token-act-pair · gpt-3.5-turbo

    instances of the phrase "it is fair to" or alternative expressions of fairness and judgment

    oai_token-act-pair · gpt-4o-miniTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res-jb
    Configuration
    jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Negative Logits
    lime
    -0.72
    listed
    -0.66
    otti
    -0.65
    calling
    -0.64
    riot
    -0.63
    pins
    -0.62
    furt
    -0.61
    lines
    -0.59
     Closing
    -0.59
    owsky
    -0.58
    POSITIVE LOGITS
     compare
    1.14
     omit
    1.10
     speculate
    1.08
     revisit
    1.05
     introduce
    1.04
     describe
    1.03
     ask
    1.03
     conclude
    1.03
     summarize
    1.03
     contemplate
    1.02
    Activations Density 0.143%

    No Known Activations