INDEX

Explanations

phrases that start with letters followed by a number, possibly indicating a section title or label

oai_token-act-pair · gpt-3.5-turbo

occurrences of the letter 'A'

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

 organisms

-0.78

âĢķ

-0.71

 talents

-0.66

—-

-0.66

Avoid

-0.66

 evils

-0.65

 endeavors

-0.64

 creeps

-0.64

frames

-0.64

 embodied

-0.63

POSITIVE LOGITS

 spokeswoman

1.43

 spokesman

1.40

 spokesperson

1.38

cknowled

1.08

gency

1.07

 representative

1.06

 statement

0.91

mber

0.89

ctors

0.89

ffect

0.88

Activations Density 0.148%

phrases that start with letters followed by a number, possibly indicating a section title or label

occurrences of the letter 'A'

No Comments

No Known Activations