INDEX

Explanations

phrases related to exaggerated qualities or possibilities

oai_token-act-pair · gpt-3.5-turbo

conditional or potential phrases indicating possibility or capability

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

 Continued

-0.88

 Spoiler

-0.75

 Emails

-0.72

 escalated

-0.70

 Cance

-0.68

chy

-0.65

 Emil

-0.65

 Lauder

-0.61

 correcting

-0.61

 reopened

-0.60

POSITIVE LOGITS

 muster

1.30

 imagine

1.14

 conceive

1.10

 think

1.00

 afford

0.99

 expect

0.97

 find

0.93

 devise

0.91

 cram

0.91

 hope

0.90

Activations Density 0.108%

phrases related to exaggerated qualities or possibilities

conditional or potential phrases indicating possibility or capability

No Comments

No Known Activations