INDEX

Explanations

references to specific numerical data or figures

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

 Mirage

-0.79

 reflex

-0.66

 vulner

-0.65

 mounts

-0.65

 conduc

-0.65

 ways

-0.65

 pleasures

-0.64

 pressures

-0.64

 Zeit

-0.63

 condem

-0.62

POSITIVE LOGITS

ï¸ı

1.22

ternity

1.02

uthor

0.94

0.93

ACP

0.90

Â§

0.88

actual

0.87

\-

0.86

âĸł

0.85

ishable

0.85

Activations Density 0.427%

references to specific numerical data or figures

No Comments

No Known Activations