INDEX

Explanations

phrases related to political and social commentary

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

 Armored

-0.56

 Dish

-0.56

 Tamil

-0.56

 Mecca

-0.56

 Heights

-0.56

ONSORED

-0.55

 Erit

-0.55

Khe

-0.54

CJ

-0.54

 scoop

-0.54

POSITIVE LOGITS

abouts

1.65

upon

1.26

after

0.95

fore

0.92

FORE

0.83

etheless

0.77

ngth

0.76

 aren

0.76

with

0.76

are

0.76

Activations Density 0.620%

phrases related to political and social commentary

No Comments

No Known Activations