INDEX

Explanations

acronyms or abbreviations including "AB"

oai_token-act-pair · gpt-3.5-turbo

references to bills or legislative actions, particularly those labeled as "AB"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 McKenna

-0.68

 Gors

-0.67

ician

-0.66

andum

-0.65

ãĥīãĥ©ãĤ´ãĥ³

-0.63

 Beir

-0.62

 Fargo

-0.62

 Daytona

-0.61

 Cannes

-0.60

 Honour

-0.60

POSITIVE LOGITS

stract

1.09

RAM

1.02

RIC

1.00

yrinth

0.92

raham

0.92

AB

0.89

STR

0.89

uilt

0.84

JECT

0.83

BA

0.83

Activations Density 0.004%

acronyms or abbreviations including "AB"

references to bills or legislative actions, particularly those labeled as "AB"

No Comments

No Known Activations