INDEX

Explanations

phrases indicating actions or intentions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "they" and phrases indicating actions or characteristics attributed to a group

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

 Starts

-0.91

 touches

-0.72

 Pieces

-0.67

 highlights

-0.66

 Samar

-0.64

 Windsor

-0.64

 leaps

-0.63

 stops

-0.62

 Highlands

-0.60

 starts

-0.60

POSITIVE LOGITS

have

1.10

want

1.07

are

1.04

rely

1.04

agree

1.02

ve

1.02

ll

0.99

re

0.99

their

0.96

react

0.95

Activations Density 0.132%

phrases indicating actions or intentions

instances of the word "they" and phrases indicating actions or characteristics attributed to a group

No Comments

No Known Activations