© 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. 8470
    Prev
    Next
    INDEX
    Explanations

    instances of the term "track record" in sentences, likely analyzing for references to past performance or reliability

    oai_token-act-pair · gpt-3.5-turbo

    references to the concept of a "track record" or history of performance

    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
    hap
    -0.75
    ishable
    -0.73
     Flavoring
    -0.72
    adin
    -0.71
    forth
    -0.71
    liga
    -0.71
     uncond
    -0.71
    essel
    -0.70
    hammad
    -0.69
    arant
    -0.66
    POSITIVE LOGITS
     resemblance
    0.69
    OA
    0.69
     record
    0.66
    making
    0.64
     Applicant
    0.64
     Acquisition
    0.63
     breaker
    0.63
     burner
    0.62
     Unlock
    0.62
     holder
    0.62
    Activations Density 0.037%

    No Known Activations