INDEX

Explanations

instances of the word "Another" used to introduce a new piece of information or detail

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

bows

-0.90

hips

-0.85

endas

-0.81

folios

-0.79

ouls

-0.79

icides

-0.79

riages

-0.78

anism

-0.78

olas

-0.78

alties

-0.77

POSITIVE LOGITS

worldly

1.06

 example

1.03

 aspect

1.01

 thing

0.98

 factor

0.97

 notable

0.97

 noteworthy

0.96

 complication

0.95

 important

0.94

 reason

0.94

Activations Density 0.054%

instances of the word "Another" used to introduce a new piece of information or detail

No Comments

No Known Activations