INDEX

Explanations

prefixes followed by hyphens

oai_token-act-pair · gpt-3.5-turbo

phrases that start with "pre-" indicating something that occurs before another event or condition

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

 Schmidt

-0.69

 DRAG

-0.63

 Stranger

-0.62

 Madd

-0.62

 Sweeney

-0.62

 Woodward

-0.62

ĸļ

-0.61

 Hyde

-0.60

 Slim

-0.60

DeV

-0.60

POSITIVE LOGITS

existing

1.21

jud

1.15

trial

1.15

release

1.14

sent

1.14

requisites

1.12

existent

1.07

season

1.06

empt

1.06

processing

1.03

Activations Density 0.024%

prefixes followed by hyphens

phrases that start with "pre-" indicating something that occurs before another event or condition

No Comments

No Known Activations