INDEX

Explanations

adverbs expressing a final outcome or conclusion

oai_token-act-pair · gpt-3.5-turbo

phrases indicating inevitability or final outcomes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

Ey

-0.71

 Frog

-0.71

AG

-0.69

 Warm

-0.67

Wee

-0.66

 Gore

-0.65

 Observer

-0.65

 Sick

-0.64

bara

-0.64

 Sung

-0.63

POSITIVE LOGITS

 ultimately

0.88

 culmin

0.87

 conduc

0.83

wcs

0.81

 Ultimately

0.80

 dismantle

0.79

Ultimately

0.77

 eventually

0.75

 succeed

0.73

 succumb

0.73

Activations Density 0.008%

adverbs expressing a final outcome or conclusion

phrases indicating inevitability or final outcomes

No Comments

No Known Activations