INDEX

Explanations

phrases indicating a selection or subset of items

oai_token-act-pair · gpt-3.5-turbo

references to specific entities or categories referred to as "ones."

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

 Membership

-0.64

Rush

-0.62

 Measures

-0.58

ãģ®å®

-0.58

Recomm

-0.58

-0.57

 cannabin

-0.57

 Payments

-0.57

 Phen

-0.56

BN

-0.56

POSITIVE LOGITS

hots

1.69

hot

1.21

selves

1.14

omething

1.13

creen

1.03

elf

0.95

paces

0.93

cale

0.92

uits

0.92

eries

0.84

Activations Density 0.035%

phrases indicating a selection or subset of items

references to specific entities or categories referred to as "ones."

No Comments

No Known Activations