INDEX

Explanations

phrases related to patents and trademarks

oai_token-act-pair · gpt-3.5-turbo

mentions of patents and trademark-related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Revel

-0.73

 Isles

-0.71

 Lans

-0.70

 Staten

-0.64

 Feast

-0.62

 Atmosp

-0.62

gaard

-0.61

 Patron

-0.61

TTC

-0.60

ashington

-0.60

POSITIVE LOGITS

 infringement

1.02

onomy

0.85

enced

0.82

ability

0.81

 examiner

0.79

 patent

0.78

 filings

0.78

ellectual

0.76

 granted

0.76

iously

0.76

Activations Density 0.029%

phrases related to patents and trademarks

mentions of patents and trademark-related terms

No Comments

No Known Activations