INDEX

Explanations

mentions of famous personalities and political figures along with negative associations

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

eem

-0.71

 Apart

-0.66

 behold

-0.60

Interested

-0.60

 Pair

-0.58

icking

-0.57

etting

-0.55

TG

-0.55

hand

-0.55

ove

-0.54

POSITIVE LOGITS

 been

1.70

been

1.44

 undergone

1.30

 gotten

1.25

 become

1.25

 begun

1.22

 Been

1.15

 risen

1.12

 gone

1.12

 arisen

1.07

Activations Density 0.803%

mentions of famous personalities and political figures along with negative associations

No Comments

No Known Activations