INDEX

Explanations

mentions of bread-related terms

oai_token-act-pair · gpt-3.5-turbo

references to bread

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

uate

-0.82

igious

-0.71

 Cheong

-0.69

OWER

-0.67

NRS

-0.66

uated

-0.66

ORT

-0.66

ARP

-0.65

ULT

-0.65

IGHT

-0.65

POSITIVE LOGITS

fruit

1.11

 Seym

1.07

 bread

1.06

grain

1.04

 dough

1.01

sticks

1.01

 pudding

1.00

winner

0.97

 oven

0.91

cake

0.90

Activations Density 0.008%

mentions of bread-related terms

references to bread

No Comments

No Known Activations