INDEX

Explanations

glyphs or unique characters in a text

oai_token-act-pair · gpt-3.5-turbo

references to numeric data and statistical information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 spitting

-0.98

 yours

-0.90

 paddle

-0.88

 shelling

-0.87

 taste

-0.87

 swoop

-0.86

 neighb

-0.86

 star

-0.84

 eleph

-0.83

swe

-0.83

POSITIVE LOGITS

CONCLUS

1.95

References

1.90

Conclusion

1.84

Figure

1.83

Introduction

1.80

FIG

1.78

Discussion

1.77

Methods

1.77

Acknowled

1.74

Thus

1.71

Activations Density 0.387%

glyphs or unique characters in a text

references to numeric data and statistical information

No Comments

No Known Activations