INDEX

Explanations

phrases related to reactions, statements, and disputes

oai_token-act-pair · gpt-3.5-turbo

phrases indicating strong reactions or emotional responses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

izon

-0.78

vre

-0.71

orn

-0.68

ocl

-0.65

grain

-0.63

herent

-0.63

scription

-0.62

que

-0.62

alogue

-0.62

enture

-0.61

POSITIVE LOGITS

 albeit

1.35

 namely

1.05

 prompting

1.02

 although

0.99

 noting

0.95

 though

0.94

 citing

0.93

 lest

0.91

 however

0.90

 according

0.88

Activations Density 0.815%

phrases related to reactions, statements, and disputes

phrases indicating strong reactions or emotional responses

No Comments

No Known Activations