INDEX

Explanations

acronyms related to legislation or organizations

oai_token-act-pair · gpt-3.5-turbo

references to the National Rifle Association (NRA) and related legislative terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

lace

-0.96

 MacArthur

-0.80

iaries

-0.78

acity

-0.77

terson

-0.76

ician

-0.75

icians

-0.72

 Kendall

-0.70

 Schwar

-0.69

iary

-0.68

POSITIVE LOGITS

VE

1.16

ISE

1.15

VEN

1.13

GE

1.06

VEL

1.05

veyard

1.05

BLE

1.00

BILITY

0.99

IN

0.96

UTH

0.95

Activations Density 0.007%

acronyms related to legislation or organizations

references to the National Rifle Association (NRA) and related legislative terms

No Comments

No Known Activations