INDEX

Explanations

the word "haven't" followed by a verb

oai_token-act-pair · gpt-3.5-turbo

instances of the auxiliary verb "have" in various forms indicating something that has not been done or completed

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 princ

-0.78

 unnecess

-0.75

anwhile

-0.71

 flies

-0.66

 descending

-0.65

 manual

-0.64

 recip

-0.63

 ingred

-0.63

 satell

-0.62

 gobl

-0.62

POSITIVE LOGITS

't

1.79

ited

1.06

ÃŃ

1.00

itely

0.92

ned

0.90

Â´

0.89

clear

0.87

ald

0.86

Õ

0.85

SPONSORED

0.85

Activations Density 0.024%

the word "haven't" followed by a verb

instances of the auxiliary verb "have" in various forms indicating something that has not been done or completed

No Comments

No Known Activations