INDEX

Explanations

proper nouns, potentially associated with places or organizations

oai_token-act-pair · gpt-3.5-turbo

proper nouns and names in the text

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

worldly

-0.75

izoph

-0.69

Quantity

-0.66

twitch

-0.65

bearing

-0.61

Evil

-0.60

ãĤ¤ãĥĪ

-0.60

 fashionable

-0.60

Ò

-0.60

Soul

-0.59

POSITIVE LOGITS

 declined

1.41

 meanwhile

1.37

 also

1.32

 spokeswoman

1.26

 spokesman

1.22

 disagreed

1.20

 echoed

1.18

 countered

1.16

 disagrees

1.11

 apologized

1.11

Activations Density 0.598%

proper nouns, potentially associated with places or organizations

proper nouns and names in the text

No Comments

No Known Activations