INDEX

Explanations

keywords related to physical weights or measurements

oai_token-act-pair · gpt-3.5-turbo

references to weight or weights in various contexts

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

 bios

-0.66

Es

-0.63

 firsthand

-0.63

 Mesa

-0.63

obs

-0.62

 counselors

-0.61

 Trib

-0.60

 scams

-0.59

En

-0.59

 Ascend

-0.59

POSITIVE LOGITS

weight

4.68

weights

3.45

Weight

2.17

 weight

1.99

 Weight

1.94

 weights

1.81

pound

1.20

 weighted

1.14

balance

1.13

 heavyweight

1.10

Activations Density 0.012%

keywords related to physical weights or measurements

references to weight or weights in various contexts

No Comments

No Known Activations