INDEX

Explanations

statements related to political opinions and beliefs

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

 Remastered

-0.72

è¦ļéĨĴ

-0.70

ĸļ

-0.66

mat

-0.65

arette

-0.64

aples

-0.63

 cellul

-0.61

abies

-0.60

ernel

-0.60

idated

-0.59

POSITIVE LOGITS

eous

1.37

 wing

1.03

wing

0.98

 fielder

0.98

 winger

0.94

 flank

0.89

move

0.86

ward

0.85

headed

0.83

 smack

0.79

Activations Density 1.053%

statements related to political opinions and beliefs

No Comments

No Known Activations