INDEX

Explanations

statements indicating a contradiction or clarification

oai_token-act-pair · gpt-3.5-turbo

the phrase "in fact."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

FTWARE

-0.85

iris

-0.77

soDeliveryDate

-0.73

-0.71

SourceFile

-0.70

asks

-0.68

ð

-0.68

kamp

-0.66

Alert

-0.66

-0.65

POSITIVE LOGITS

 essence

1.27

 fact

1.24

 theory

1.21

 effect

1.20

 turn

1.12

 principle

1.11

 aggregate

1.02

 retrospect

1.01

 hindsight

0.97

 reality

0.93

Activations Density 0.071%

statements indicating a contradiction or clarification

the phrase "in fact."

No Comments

No Known Activations