INDEX

Explanations

phrases with the word "take"

oai_token-act-pair · gpt-3.5-turbo

expressions of taking actions or holding views

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

hoff

-0.73

atron

-0.70

dropping

-0.66

fold

-0.65

Haunted

-0.65

idding

-0.63

destroy

-0.63

 Rouge

-0.62

band

-0.61

éŃĶ

-0.59

POSITIVE LOGITS

 advantage

1.12

 aback

1.03

 kindly

1.00

 heed

0.98

 cues

0.96

 precedence

0.94

aways

0.92

 precautions

0.91

 seriously

0.90

 cogn

0.89

Activations Density 0.088%

phrases with the word "take"

expressions of taking actions or holding views

No Comments

No Known Activations