INDEX

Explanations

news headlines or articles related to political events or controversies

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

 deciding

-0.63

ãĢĲ

-0.62

stood

-0.62

represented

-0.61

profits

-0.59

 SHALL

-0.59

FTWARE

-0.59

abiding

-0.59

CVE

-0.59

itatively

-0.58

POSITIVE LOGITS

 IMAGES

1.15

utterstock

1.04

 Shutterstock

0.97

ulton

0.95

photos

0.93

 PRESS

0.89

 Photos

0.86

 Images

0.84

 Photography

0.84

Photo

0.82

Activations Density 1.214%

news headlines or articles related to political events or controversies

No Comments

No Known Activations