INDEX

Explanations

phrases or acronyms related to specific events or organizations, particularly those ending in "CS" with varying degrees of specificities

oai_token-act-pair · gpt-3.5-turbo

references to a specific FDA regulatory classification or category

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

 conspicuous

-0.65

judicial

-0.65

 sale

-0.65

accompanied

-0.65

 maiden

-0.64

 wills

-0.64

life

-0.64

estate

-0.63

 persu

-0.63

accompan

-0.63

POSITIVE LOGITS

CS

1.00

IRO

1.00

ANE

0.99

DF

0.96

OME

0.95

DS

0.95

DP

0.89

ENSE

0.87

GN

0.86

ULAR

0.85

Activations Density 0.006%

phrases or acronyms related to specific events or organizations, particularly those ending in "CS" with varying degrees of specificities

references to a specific FDA regulatory classification or category

No Comments

No Known Activations