INDEX

Explanations

phrases starting with "We"

oai_token-act-pair · gpt-3.5-turbo

instances of the word "We."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 reflex

-0.69

 solo

-0.62

 arch

-0.62

 minor

-0.60

 career

-0.60

 caps

-0.60

 rookie

-0.59

his

-0.59

 nose

-0.58

 lingering

-0.58

POSITIVE LOGITS

We

2.92

Our

2.45

We

1.97

we

1.93

You

1.54

WE

1.53

Our

1.53

They

1.50

Let

1.49

There

1.47

Activations Density 0.049%

phrases starting with "We"

instances of the word "We."

No Comments

No Known Activations