INDEX

Explanations

words related to organizations and institutions, such as acronyms and names

oai_token-act-pair · gpt-3.5-turbo

references to organizations and agencies

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

wagen

-0.78

hood

-0.76

abilia

-0.72

icles

-0.71

makers

-0.68

icle

-0.67

uits

-0.65

ogue

-0.63

 Sabbath

-0.62

phrase

-0.62

POSITIVE LOGITS

DEF

0.97

 Headquarters

0.89

IL

0.89

senal

0.88

 accredited

0.84

ORN

0.80

ILA

0.80

CA

0.78

REC

0.78

OC

0.78

Activations Density 0.116%

words related to organizations and institutions, such as acronyms and names

references to organizations and agencies

No Comments

No Known Activations