INDEX

Explanations

locations or landmarks

oai_token-act-pair · gpt-3.5-turbo

proper nouns related to locations and geographic features

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

quickShipAvailable

-0.90

uably

-0.79

earcher

-0.78

ocrates

-0.77

matched

-0.75

isconsin

-0.75

ourning

-0.74

iversal

-0.72

perse

-0.72

recent

-0.71

POSITIVE LOGITS

 Estate

1.15

 Mountains

1.03

 Massacre

1.01

 Shopping

1.00

 Peninsula

0.99

 Bridge

0.98

 Express

0.98

 Corridor

0.97

 Basin

0.97

 Islands

0.96

Activations Density 0.141%

locations or landmarks

proper nouns related to locations and geographic features

No Comments

No Known Activations