INDEX

Explanations

information related to news articles and videos

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

vironment

-0.79

abil

-0.67

affles

-0.67

yss

-0.66

urance

-0.66

hap

-0.66

insula

-0.65

agall

-0.64

essee

-0.63

igate

-0.63

POSITIVE LOGITS

 clip

0.86

 footage

0.81

 Thumbnails

0.78

 clips

0.74

 Transcript

0.71

 snippet

0.71

 Surveillance

0.68

embed

0.67

 WATCHED

0.66

GAME

0.65

Activations Density 0.029%

information related to news articles and videos

No Comments

No Known Activations