INDEX

Explanations

mentions of specific details, such as lists, instructions, and versions related to various subjects

oai_token-act-pair · gpt-3.5-turbo

information related to books and their details

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

thinking

-0.74

ordinate

-0.70

 fearing

-0.68

uristic

-0.68

 disadvant

-0.64

aeda

-0.62

 WATCHED

-0.61

 Learns

-0.60

maxwell

-0.59

dq

-0.58

POSITIVE LOGITS

can

1.30

 follows

1.03

 airs

1.02

 HERE

0.94

 awaits

0.92

 will

0.91

 below

0.91

 includes

0.87

 reads

0.87

 resides

0.86

Activations Density 0.299%

mentions of specific details, such as lists, instructions, and versions related to various subjects

information related to books and their details

No Comments

No Known Activations