INDEX

Explanations

phrases or words referring to variety or differences

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

Data Type

torch.float32

Hook Point

blocks.8.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

 âĨĳ

-0.68

WI

-0.67

WARD

-0.64

HP

-0.62

OIL

-0.61

âĢł

-0.60

Never

-0.60

AFTA

-0.58

Trivia

-0.57

 Wilde

-0.56

POSITIVE LOGITS

iating

1.83

iates

1.60

iator

1.37

iations

1.37

ials

1.36

iated

1.22

 kinds

1.18

iable

1.11

iation

1.08

iate

1.07

Activations Density 0.436%

phrases or words referring to variety or differences

No Comments

No Known Activations