INDEX

Explanations

answers or responses to questions

oai_token-act-pair · gpt-3.5-turbo

affirmations or definitive answers to questions posed within the text

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.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

¥µ

-0.76

ivities

-0.62

 Orient

-0.61

apons

-0.60

idi

-0.60

 hobbies

-0.60

ombat

-0.59

iannopoulos

-0.58

spir

-0.58

 activities

-0.57

POSITIVE LOGITS

YES

1.13

yes

1.08

YES

1.01

yes

0.94

 affirmative

0.86

 Nope

0.84

Yes

0.80

nil

0.78

answer

0.76

brainer

0.74

Activations Density 0.148%

answers or responses to questions

affirmations or definitive answers to questions posed within the text

No Comments

No Known Activations