INDEX

Explanations

phrases related to direct speech and thoughts

oai_token-act-pair · gpt-3.5-turbo

repeated instances of the word "I" and variations that indicate uncertainty or negative assertions

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

 Palest

-0.76

anwhile

-0.71

 Mous

-0.69

Zup

-0.63

 Manhattan

-0.62

 recogn

-0.61

 Franch

-0.60

NEC

-0.59

çĶ°

-0.59

 sights

-0.58

POSITIVE LOGITS

¡

1.05

Ķ

1.01

ľ

1.00

¤

1.00

Ń

1.00

ĺ

0.96

âĢķ

0.96

¬

0.95

¢

0.95

Ļ

0.95

Activations Density 0.297%

phrases related to direct speech and thoughts

repeated instances of the word "I" and variations that indicate uncertainty or negative assertions

No Comments

No Known Activations