INDEX

Explanations

verbs related to actions of taking, such as "takes," "took," "taking," and "taken."

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

lite

-0.69

raid

-0.69

è¦ļéĨĴ

-0.67

 ILCS

-0.65

gian

-0.64

eous

-0.62

lex

-0.62

 Conclusion

-0.60

arie

-0.60

 constitu

-0.60

POSITIVE LOGITS

 advantage

1.16

 refuge

0.96

 pains

0.95

aim

0.94

um

0.85

 charge

0.83

 place

0.82

 part

0.82

 inspiration

0.82

 aback

0.81

Activations Density 0.074%

verbs related to actions of taking, such as "takes," "took," "taking," and "taken."

No Comments

No Known Activations