INDEX

Explanations

the phrase 'for good'

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

olate

-0.81

ulous

-0.68

otom

-0.63

uine

-0.63

enged

-0.62

uten

-0.62

gerald

-0.62

ibel

-0.61

peak

-0.60

yss

-0.59

POSITIVE LOGITS

 reason

1.34

 reasons

1.28

 measure

1.10

 Reasons

0.95

 purposes

0.93

 sake

0.93

 Reason

0.86

 measures

0.81

reason

0.76

 cause

0.70

Activations Density 0.032%

the phrase 'for good'

No Comments

No Known Activations