INDEX

Explanations

phrases related to arguments and disagreements

oai_token-act-pair · gpt-3.5-turbo

phrases related to disagreements or debates

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

ionage

-0.83

ngth

-0.82

iaries

-0.81

vik

-0.72

OGR

-0.72

ortun

-0.69

ROR

-0.68

Ĳ

-0.68

--+

-0.67

IENCE

-0.66

POSITIVE LOGITS

 whether

1.08

 legality

0.83

 semantics

0.80

how

0.79

whether

0.78

 merits

0.78

 topics

0.71

 definitions

0.71

 footing

0.71

 nuances

0.70

Activations Density 0.185%

phrases related to arguments and disagreements

phrases related to disagreements or debates

No Comments

No Known Activations