INDEX

Explanations

prepositions followed by a noun or a phrase that is likely to be a title or subtitle

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

azines

-0.69

adelphia

-0.68

auga

-0.66

pps

-0.65

 Thousand

-0.65

wives

-0.65

 Solitaire

-0.64

Enlarge

-0.63

\<

-0.63

 Guides

-0.63

POSITIVE LOGITS

 contention

1.03

 weakness

0.94

 desperation

0.93

 persuasion

0.91

 differentiation

0.84

 pride

0.84

 excellence

0.84

 oppression

0.84

 accomplishment

0.83

 shame

0.83

Activations Density 1.003%

prepositions followed by a noun or a phrase that is likely to be a title or subtitle

No Comments

No Known Activations