INDEX

Explanations

mentions of specific newspapers and their articles

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

ardless

-0.76

sed

-0.76

venge

-0.75

cles

-0.72

iago

-0.69

isable

-0.67

hra

-0.67

pired

-0.66

utical

-0.65

artifacts

-0.65

POSITIVE LOGITS

 Magazine

1.26

 Editorial

1.11

 editorial

1.08

 columnist

1.02

 bestselling

0.96

 reporter

0.96

 headline

0.94

 article

0.93

 magazine

0.93

Magazine

0.91

Activations Density 0.083%

mentions of specific newspapers and their articles

No Comments

No Known Activations