INDEX

Explanations

organizations or entities related to reports

oai_token-act-pair · gpt-3.5-turbo

references to sources or authors in reports

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãĤ¦ãĤ¹

-0.74

uality

-0.71

chwitz

-0.71

 "$:/

-0.67

ettes

-0.67

 brakes

-0.66

rats

-0.64

atics

-0.64

iencies

-0.63

 chloride

-0.62

POSITIVE LOGITS

 virtue

0.86

products

0.84

laws

0.76

product

0.74

 Forbes

0.70

pass

0.70

 Newsweek

0.69

 Wikileaks

0.67

gone

0.67

 Uriel

0.67

Activations Density 0.102%

organizations or entities related to reports

references to sources or authors in reports

No Comments

No Known Activations