INDEX

Explanations

quotations

oai_token-act-pair · gpt-3.5-turbo

quotation marks

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 arch

-0.78

 grammar

-0.65

 appro

-0.65

 seasoned

-0.64

 finished

-0.63

 repro

-0.63

 grades

-0.62

 overlook

-0.62

 filib

-0.62

 lamp

-0.61

POSITIVE LOGITS

We

1.21

Our

1.15

They

1.06

There

1.05

It

1.05

1.03

Everything

1.02

Today

1.02

What

1.01

Operation

0.99

Activations Density 0.106%

quotations

quotation marks

No Comments

No Known Activations