INDEX

Explanations

punctuation marks

oai_token-act-pair · gpt-3.5-turbo

phrases indicating product recommendations or performance assessments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 volunteering

-0.87

 volunteer

-0.82

 parish

-0.82

 Stew

-0.82

 pupp

-0.80

 coordinated

-0.79

 volunteers

-0.78

 drafts

-0.78

 amnesty

-0.77

 congreg

-0.77

POSITIVE LOGITS

Features

1.71

Pros

1.49

Product

1.47

Hardware

1.46

Performance

1.44

Availability

1.42

Price

1.36

Spec

1.36

Compared

1.35

Manufact

1.35

Activations Density 0.353%

punctuation marks

phrases indicating product recommendations or performance assessments

No Comments

No Known Activations