INDEX

Explanations

phrases including the words "they're"

oai_token-act-pair · gpt-3.5-turbo

references to people or groups and their opinions or actions

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

 Windsor

-0.69

 Catalyst

-0.67

 Starts

-0.65

 Tasman

-0.65

 Garmin

-0.63

 Suzuki

-0.62

CY

-0.62

 Photographer

-0.61

 Pascal

-0.60

 Samar

-0.60

POSITIVE LOGITS

were

1.19

selves

1.10

have

1.03

 themselves

1.00

are

0.99

 selves

0.91

 deserve

0.90

tarians

0.87

hip

0.86

ought

0.86

Activations Density 0.210%

phrases including the words "they're"

references to people or groups and their opinions or actions

No Comments

No Known Activations