INDEX

Explanations

mentions of the word "Trump."

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the word "trumps" and other related variations

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

raq

-0.84

ANC

-0.84

GAN

-0.79

ANCE

-0.77

zai

-0.76

 srfAttach

-0.75

RW

-0.72

ANA

-0.71

BIL

-0.70

WORK

-0.70

POSITIVE LOGITS

hift

1.00

manship

1.00

paces

0.95

pace

0.87

ilver

0.85

peed

0.84

poons

0.82

oulos

0.82

uits

0.79

hops

0.79

Activations Density 0.028%

mentions of the word "Trump."

repeated mentions of the word "trumps" and other related variations

No Comments

No Known Activations