INDEX

Explanations

numerical information or statistics presented in a tabular format

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

conservancy

-0.87

 corrid

-0.86

 undermin

-0.79

.}

-0.74

 reluct

-0.72

atonin

-0.71

soDeliveryDate

-0.69

 neighb

-0.68

 destro

-0.68

uca

-0.68

POSITIVE LOGITS

 Total

1.01

 âĵĺ

1.01

 Female

1.01

 Male

1.01

Yes

0.98

 Difference

0.97

 Notes

0.97

Avg

0.95

 None

0.91

 Unsure

0.91

Activations Density 1.513%

numerical information or statistics presented in a tabular format

No Comments

No Known Activations