INDEX

Explanations

references to religious texts and teachings

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

emouth

-0.75

 Beckham

-0.74

leneck

-0.71

 Velocity

-0.70

 Steelers

-0.68

isode

-0.68

 Nets

-0.67

 interns

-0.66

 Beaver

-0.66

 undercover

-0.65

POSITIVE LOGITS

 teachings

1.25

 scriptures

1.17

 traditions

1.06

 doctrines

1.01

 Scriptures

0.98

 theology

0.97

 scripture

0.96

 teaches

0.95

 practiced

0.94

 sect

0.93

Activations Density 0.160%

references to religious texts and teachings

No Comments

No Known Activations