INDEX

Explanations

legalese and policy-related terms

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

 thence

-0.93

ĺħ

-0.83

dl

-0.76

DragonMagazine

-0.72

Fal

-0.70

aline

-0.65

here

-0.65

ÅĤ

-0.65

çĶ

-0.64

igate

-0.63

POSITIVE LOGITS

ties

1.55

 kinds

1.47

 types

1.39

 aspects

1.30

 wavelengths

1.13

 segments

1.11

 sections

1.09

 combinations

1.09

 situations

1.06

 categories

1.05

Activations Density 0.028%

legalese and policy-related terms

No Comments

No Known Activations