INDEX

Explanations

suggestions or recommendations

oai_token-act-pair · gpt-3.5-turbo

phrases that convey recommendations or assertions

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

 Lives

-0.72

Bee

-0.68

 Bree

-0.63

 Gazette

-0.58

 Patri

-0.55

fighting

-0.55

 ILCS

-0.55

 Hall

-0.54

aign

-0.54

itude

-0.54

POSITIVE LOGITS

 suggest

3.28

 suggests

2.22

 imply

2.04

suggest

2.03

 indicate

1.94

 suggesting

1.86

 recommend

1.85

 sugg

1.80

uggest

1.80

 suggested

1.70

Activations Density 0.017%

suggestions or recommendations

phrases that convey recommendations or assertions

No Comments

No Known Activations