INDEX

Explanations

mentions of news media outlets and reports

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

inho

-0.77

 Reson

-0.69

ayne

-0.69

pired

-0.66

 Wade

-0.65

BuyableInstoreAndOnline

-0.63

vasive

-0.62

ensions

-0.62

 Stark

-0.60

 Stupid

-0.60

POSITIVE LOGITS

room

1.00

reader

0.94

orial

0.93

 coverage

0.91

 Coverage

0.91

 headlines

0.90

eval

0.89

Reporting

0.89

flash

0.88

 outlets

0.86

Activations Density 3.691%

mentions of news media outlets and reports

No Comments

No Known Activations