INDEX

Explanations

capitalized words related to governance and political figures

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

oft

-0.84

DragonMagazine

-0.83

Aren

-0.83

ASC

-0.81

-0.80

-0.79

Reloaded

-0.79

chev

-0.79

Thumbnail

-0.78

usc

-0.77

POSITIVE LOGITS

 Â¯

0.88

0.87

 tion

0.77

0.75

 SECTION

0.73

0.70

0.67

0.65

0.64

ITION

0.64

Activations Density 0.046%

capitalized words related to governance and political figures

No Comments

No Known Activations