INDEX

Explanations

phrases related to financial penalties

oai_token-act-pair · gpt-3.5-turbo

instances of penalties or fines

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

leader

-0.75

NAS

-0.66

/(

-0.65

 Ezek

-0.65

 Headquarters

-0.64

ney

-0.61

 Izan

-0.61

 confronts

-0.60

seekers

-0.60

boro

-0.60

POSITIVE LOGITS

 fine

3.91

fine

2.54

Fine

2.30

 Fine

2.20

 fines

1.61

 finer

1.60

 alright

1.55

 okay

1.51

OK

1.41

ok

1.40

Activations Density 0.023%

phrases related to financial penalties

instances of penalties or fines

No Comments

No Known Activations