INDEX

Explanations

phrases summarizing information or opinions

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate summarization or conclusion

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

 bakery

-0.65

 proxies

-0.64

imer

-0.62

 promoter

-0.60

 hired

-0.59

 tracker

-0.59

 harb

-0.58

 hire

-0.58

ially

-0.57

 prototypes

-0.57

POSITIVE LOGITS

 succinct

0.92

DOS

0.83

 concise

0.76

ĺħ

0.76

 summarizes

0.74

terday

0.70

Ĥª

0.68

Â¢

0.67

 summed

0.66

NYSE

0.66

Activations Density 0.066%

phrases summarizing information or opinions

phrases that indicate summarization or conclusion

No Comments

No Known Activations