INDEX

Explanations

instructions or guidelines related to specific activities or tasks

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 expecting

-0.72

tons

-0.65

 hoping

-0.64

 forgetting

-0.64

furt

-0.63

seeing

-0.63

 advising

-0.63

 afraid

-0.62

soDeliveryDate

-0.61

 cursing

-0.60

POSITIVE LOGITS

 occur

1.31

 become

1.24

 propagate

1.21

 explode

1.16

be

1.16

 arrive

1.12

 accumulate

1.11

 originate

1.11

 evolve

1.09

 arise

1.08

Activations Density 0.184%

instructions or guidelines related to specific activities or tasks

No Comments

No Known Activations