INDEX

Explanations

phrases related to accomplishments and achievements

oai_token-act-pair · gpt-3.5-turbo

references to actions and achievements related to gratitude and contributions

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

furt

-0.61

 Burns

-0.60

inho

-0.56

arget

-0.55

 faint

-0.54

anchester

-0.53

Reilly

-0.52

direction

-0.52

 smokes

-0.51

eware

-0.50

POSITIVE LOGITS

 since

1.32

 lately

1.24

 thus

1.06

 hitherto

1.04

since

0.96

 historically

0.91

 previously

0.85

 recently

0.85

 throughout

0.83

 over

0.80

Activations Density 0.315%

phrases related to accomplishments and achievements

references to actions and achievements related to gratitude and contributions

No Comments

No Known Activations