INDEX

Explanations

phrases related to disputes or disagreements

oai_token-act-pair · gpt-3.5-turbo

references to points of disagreement or dispute

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

hen

-0.79

 Patron

-0.71

aper

-0.70

 Spells

-0.68

RN

-0.64

hex

-0.64

ghan

-0.64

EH

-0.62

ge

-0.60

 dial

-0.60

POSITIVE LOGITS

 contention

4.20

 assertion

1.45

 contended

1.30

 disagreement

1.25

 conjecture

1.24

 obscurity

1.20

 assertions

1.20

 contend

1.15

 allegation

1.12

 inconsistency

1.10

Activations Density 0.022%

phrases related to disputes or disagreements

references to points of disagreement or dispute

No Comments

No Known Activations