INDEX

Explanations

words related to stories or news events

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

izens

-0.75

tsky

-0.75

ivities

-0.71

enhagen

-0.70

nings

-0.68

eways

-0.68

cuts

-0.68

heast

-0.67

gow

-0.66

tz

-0.66

POSITIVE LOGITS

 albeit

1.36

 akin

1.11

albeit

1.06

 whereas

1.03

 although

1.01

 requiring

1.01

 meaning

0.97

 hence

0.96

 consisting

0.95

 specializing

0.95

Activations Density 0.275%

words related to stories or news events

No Comments

No Known Activations

words related to stories or news events

No Comments

No Known Activations