INDEX

Explanations

references to historical events or political statements

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

ebin

-0.90

aspx

-0.84

 waiver

-0.77

Alert

-0.76

Tickets

-0.75

 waivers

-0.74

Quality

-0.73

 blocker

-0.72

 referral

-0.70

Applic

-0.69

POSITIVE LOGITS

era

1.40

Era

1.39

 eras

1.21

oslov

1.14

 centuries

1.14

 Yugoslavia

1.14

1.13

 civilizations

1.12

1.11

1.09

Activations Density 7.435%

references to historical events or political statements

No Comments

No Known Activations