INDEX

Explanations

phrases containing the letters "ab"

oai_token-act-pair · gpt-3.5-turbo

the prefix "ab-" indicating a negative or reverse quality

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Marian

-0.67

 Cups

-0.65

 attorneys

-0.65

Rex

-0.61

 Attorney

-0.60

 attorney

-0.59

 offshore

-0.59

 Hendricks

-0.58

 virtue

-0.58

 Maker

-0.57

POSITIVE LOGITS

omination

1.24

rog

1.14

ject

1.08

rid

1.07

ounding

1.03

omin

1.01

raham

1.01

err

0.99

acus

0.98

usable

0.98

Activations Density 0.017%

phrases containing the letters "ab"

the prefix "ab-" indicating a negative or reverse quality

No Comments

No Known Activations