INDEX

Explanations

phrases related to intention or potential actions

oai_token-act-pair · gpt-3.5-turbo

the verb "be" in various constructions implying existence or conditions

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rather

-0.63

maybe

-0.63

strous

-0.63

 Might

-0.62

 guise

-0.62

 remind

-0.62

 MUST

-0.61

MAY

-0.61

not

-0.60

staking

-0.59

POSITIVE LOGITS

 able

1.21

 anymore

1.16

 bothered

1.01

 tolerated

0.96

 harmed

0.92

 swayed

0.89

icable

0.88

 counted

0.87

 anywhere

0.86

 allowed

0.84

Activations Density 0.107%

phrases related to intention or potential actions

the verb "be" in various constructions implying existence or conditions

No Comments

No Known Activations