INDEX

Explanations

references to specific chapters or sections

oai_token-act-pair · gpt-3.5-turbo

references to chapters or episodes in a narrative context

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 claim

-0.77

 sling

-0.72

 tweet

-0.71

 optics

-0.68

 lineup

-0.67

 Britt

-0.66

 choice

-0.66

gem

-0.66

 tweeted

-0.66

 shooter

-0.65

POSITIVE LOGITS

Chapter

4.07

CHAPTER

2.47

 Chapter

1.85

chapter

1.67

APTER

1.60

 Chapters

1.55

Volume

1.46

Episode

1.45

Author

1.42

apters

1.39

Activations Density 0.028%

references to specific chapters or sections

references to chapters or episodes in a narrative context

No Comments

No Known Activations