INDEX

Explanations

references to specific names of people, places, or organizations

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

/"

-0.65

 alongside

-0.60

—"

-0.59

Ãĥ

-0.58

alone

-0.57

 beforehand

-0.57

—-

-0.57

omever

-0.56

iod

-0.56

egu

-0.55

POSITIVE LOGITS

odore

1.00

resa

1.00

atre

0.92

 longest

0.80

ater

0.80

 fastest

0.79

orem

0.78

 greatest

0.78

 Latest

0.78

 largest

0.78

Activations Density 0.603%

references to specific names of people, places, or organizations

No Comments

No Known Activations