INDEX

Explanations

mentions of the word "other."

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

zon

-0.72

iste

-0.71

veland

-0.66

ãĥĥ

-0.66

zai

-0.64

gain

-0.63

_-

-0.62

 Retrieved

-0.61

Progress

-0.60

abad

-0.60

POSITIVE LOGITS

worldly

1.79

 side

1.05

 senses

0.80

world

0.78

 aspect

0.77

 thing

0.77

 half

0.76

 ones

0.75

iest

0.73

 aspects

0.73

Activations Density 0.043%

mentions of the word "other."

No Comments

No Known Activations