INDEX

Explanations

instances of the verb "was" followed by adjectives or prepositional phrases

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the word “was.”

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

 Individuals

-0.67

 Current

-0.66

 preserves

-0.63

 Which

-0.62

 Highlands

-0.62

 entails

-0.61

 Various

-0.61

 Countries

-0.60

 Replacement

-0.60

bies

-0.60

POSITIVE LOGITS

wolves

1.01

 fortunate

0.99

 hoping

0.97

 wondering

0.97

 able

0.96

 expecting

0.96

 tempted

0.92

 unable

0.92

 lucky

0.91

 afraid

0.91

Activations Density 0.201%

instances of the verb "was" followed by adjectives or prepositional phrases

repeated mentions of the word “was.”

No Comments

No Known Activations