INDEX

Explanations

words and phrases related to access or permission

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

 Kern

-0.86

 farewell

-0.62

Maker

-0.62

 abst

-0.59

 estranged

-0.59

 bulls

-0.58

 disband

-0.58

 circling

-0.56

 almond

-0.55

birds

-0.55

POSITIVE LOGITS

ibility

1.46

ibly

1.35

ible

1.23

ory

1.00

ibles

1.00

ibl

0.98

ioned

0.95

ories

0.95

IBLE

0.94

ibilities

0.93

Activations Density 0.548%

words and phrases related to access or permission

No Comments

No Known Activations