INDEX

Explanations

targeted keywords related to specific techniques or methods

oai_token-act-pair · gpt-3.5-turbo

references to various methods or techniques

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 impression

-0.65

 earnest

-0.64

 Commodore

-0.63

 Aliens

-0.62

 gloom

-0.61

 juven

-0.61

 Tyrann

-0.60

 Sons

-0.60

ternity

-0.59

 Messi

-0.59

POSITIVE LOGITS

ologies

1.88

ologically

1.41

ical

1.27

ology

1.27

ologic

1.26

ically

1.20

ological

1.11

OLOGY

1.10

ologists

1.10

ologist

1.03

Activations Density 0.024%

targeted keywords related to specific techniques or methods

references to various methods or techniques

No Comments

No Known Activations