INDEX

Explanations

phrases related to historical processes or events

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

istan

-0.69

gon

-0.67

uckle

-0.65

 CARD

-0.62

gang

-0.60

raid

-0.59

vest

-0.58

ves

-0.58

meet

-0.58

urden

-0.56

POSITIVE LOGITS

place

1.19

body

1.18

how

1.15

 semblance

1.10

ones

1.10

 sort

0.86

HOW

0.83

 notable

0.79

ONE

0.79

 unspecified

0.78

Activations Density 0.493%

phrases related to historical processes or events

No Comments

No Known Activations