INDEX

Explanations

phrases related to personal or anecdotal stories

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

 mell

-0.66

 Gerry

-0.62

 congr

-0.61

 newsletters

-0.60

lihood

-0.60

 frequent

-0.59

 Suns

-0.59

 balloons

-0.58

 Compton

-0.58

ATED

-0.58

POSITIVE LOGITS

eless

1.33

rock

1.25

blers

1.24

aji

1.10

urai

1.00

ble

0.97

ming

0.94

arak

0.93

ash

0.91

okin

0.91

Activations Density 0.018%

phrases related to personal or anecdotal stories

No Comments

No Known Activations