INDEX

Explanations

phrases related to political leadership or representation

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

 territ

-0.65

 challeng

-0.62

aimon

-0.61

YS

-0.61

 reper

-0.59

ÏĢ

-0.58

ske

-0.57

USE

-0.56

 deaf

-0.55

 tandem

-0.54

POSITIVE LOGITS

ing

1.96

ed

1.58

ership

1.26

ington

1.14

quartered

1.13

ions

1.12

edly

1.11

ers

1.08

edIn

1.07

ments

1.05

Activations Density 0.176%

phrases related to political leadership or representation

No Comments

No Known Activations