INDEX

Explanations

instances where the word "a" is followed by certain context-specific terms

oai_token-act-pair · gpt-3.5-turbo

articles and determiners that indicate nouns

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Arcade

-0.76

 Extras

-0.73

antry

-0.69

scenes

-0.69

enda

-0.67

scripts

-0.66

abilities

-0.64

ographs

-0.64

Links

-0.63

insk

-0.63

POSITIVE LOGITS

lot

1.23

 person

1.15

cknowled

1.01

chie

1.01

 majority

1.00

rouse

0.99

cknow

0.99

 lack

0.99

 politician

0.93

verages

0.91

Activations Density 0.235%

instances where the word "a" is followed by certain context-specific terms

articles and determiners that indicate nouns

No Comments

No Known Activations