INDEX

Explanations

words related to being captivated or captivating

oai_token-act-pair · gpt-3.5-turbo

forms of the word "captivate" and related terms

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

sterdam

-0.70

 Integrity

-0.68

 partnership

-0.67

 Workers

-0.65

 Enterprise

-0.65

 trust

-0.64

 Worker

-0.64

ukemia

-0.64

ãĤ§

-0.64

Pax

-0.63

POSITIVE LOGITS

ivating

1.12

ivated

1.02

aining

1.01

ained

1.00

anted

0.99

icol

0.94

aning

0.88

anting

0.86

ains

0.86

iors

0.86

Activations Density 0.017%

words related to being captivated or captivating

forms of the word "captivate" and related terms

No Comments

No Known Activations