INDEX

Explanations

emotions or states related to interactions or situations

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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

artifacts

-0.76

è£ħ

-0.73

 dated

-0.72

Downloadha

-0.71

Unique

-0.71

otide

-0.70

JO

-0.69

plane

-0.68

unique

-0.68

POSITIVE LOGITS

ness

1.38

 onlook

1.20

NESS

1.00

 individuals

0.98

 souls

0.96

 masses

0.96

 shoppers

0.96

 populace

0.95

 youngsters

0.94

nesses

0.92

Activations Density 0.337%

emotions or states related to interactions or situations

No Comments

No Known Activations