INDEX

Explanations

phrases related to activism or protests

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

inguishable

-0.73

onom

-0.70

ongo

-0.66

imet

-0.65

PDF

-0.65

abul

-0.64

eport

-0.63

velop

-0.63

omal

-0.63

utsche

-0.61

POSITIVE LOGITS

 blah

1.22

 secondly

1.13

etc

1.12

 furthermore

1.02

 lest

0.99

 THEN

0.96

oh

0.93

 thereby

0.88

etc

0.88

)).

0.85

Activations Density 0.603%

phrases related to activism or protests

No Comments

No Known Activations