INDEX

Explanations

safety-related instructions or warnings

oai_token-act-pair · gpt-3.5-turbo

warnings and advisories related to safety

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

 triumphant

-0.62

Rank

-0.61

ãĤ´

-0.57

Winner

-0.57

 unemployment

-0.57

 Achievement

-0.57

Spons

-0.57

 losers

-0.56

 Unemployment

-0.56

 Difference

-0.55

POSITIVE LOGITS

 beware

1.71

 caution

1.52

eware

1.37

 precautions

1.34

avoid

1.33

Avoid

1.26

 avoid

1.25

 precaution

1.25

 Beware

1.24

 Avoid

1.21

Activations Density 0.444%

safety-related instructions or warnings

warnings and advisories related to safety

No Comments

No Known Activations