INDEX

Explanations

mentions of political figures and events in the context of government policies and international relations

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

replace

-0.81

gat

-0.79

 namely

-0.76

rand

-0.72

craft

-0.71

ftime

-0.70

thood

-0.68

watching

-0.68

-0.67

.--

-0.66

POSITIVE LOGITS

 entire

1.51

 entirety

1.45

 remainder

1.35

 slightest

1.29

 same

1.24

 whole

1.19

 latter

1.17

ses

1.16

 smallest

1.15

 brunt

1.13

Activations Density 1.546%

mentions of political figures and events in the context of government policies and international relations

No Comments

No Known Activations