INDEX

Explanations

text containing the word "can" followed by other words or phrases

oai_token-act-pair · gpt-3.5-turbo

the phrase "you can" and its variations

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

Moz

-0.65

burgh

-0.65

 Mant

-0.64

é¾įå¥ĳå£«

-0.62

 revision

-0.61

 Gateway

-0.61

 Yards

-0.59

 scrapped

-0.58

 Strikes

-0.58

 Sandwich

-0.58

POSITIVE LOGITS

't

1.51

NOT

1.12

berra

0.97

adian

0.95

isters

0.94

 easily

0.89

ister

0.87

 customize

0.86

 optionally

0.86

 choose

0.85

Activations Density 0.106%

text containing the word "can" followed by other words or phrases

the phrase "you can" and its variations

No Comments

No Known Activations