INDEX

Explanations

words related to negative or harmful situations and conditions

oai_token-act-pair · gpt-3.5-turbo

terms related to malpractice

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

loo

-0.84

 Carbuncle

-0.73

 Teach

-0.73

externalActionCode

-0.67

 Statements

-0.66

 Emirates

-0.66

 Annotations

-0.66

 Hobby

-0.66

 Polo

-0.65

DragonMagazine

-0.65

POSITIVE LOGITS

ignant

1.05

colm

1.02

adies

0.98

igned

0.98

formed

0.96

practice

0.95

absor

0.93

ef

0.91

cellaneous

0.90

igning

0.85

Activations Density 0.024%

words related to negative or harmful situations and conditions

terms related to malpractice

No Comments

No Known Activations