INDEX

Explanations

phrases related to legal and political news

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

 unemploy

-0.81

reon

-0.70

 nurt

-0.67

$$

-0.66

 lett

-0.65

 combatants

-0.64

 spoil

-0.64

 creeps

-0.63

 looms

-0.62

 RIGHT

-0.61

POSITIVE LOGITS

cluding

1.28

 addition

1.27

spect

1.22

 contrast

1.20

ventory

1.20

jured

1.14

 spite

1.12

vasive

1.12

cluded

1.11

mates

1.09

Activations Density 0.129%

phrases related to legal and political news

No Comments

No Known Activations