INDEX

Explanations

statements or phrases clarifying or emphasizing a point within a text

oai_token-act-pair · gpt-3.5-turbo

phrases that express certainty or conclusions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 conflic

-0.82

¥ŀ

-0.76

Tai

-0.70

xtap

-0.69

 hemor

-0.67

 notor

-0.67

itially

-0.66

 cumbers

-0.65

estern

-0.65

phrine

-0.64

POSITIVE LOGITS

 goodbye

1.46

 hello

1.06

 Goodbye

0.90

ieu

0.88

 aloud

0.87

 sorry

0.83

 farewell

0.81

bye

0.78

hi

0.73

ings

0.71

Activations Density 0.048%

statements or phrases clarifying or emphasizing a point within a text

phrases that express certainty or conclusions

No Comments

No Known Activations