INDEX

Explanations

government or law-related entities or positions

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

isable

-0.73

olini

-0.67

oise

-0.67

Interstitial

-0.67

 Passing

-0.66

 esteem

-0.64

obyl

-0.64

isers

-0.63

ipedia

-0.63

POSITIVE LOGITS

 prosec

1.03

holder

0.84

 clerk

0.81

holders

0.80

 Cheong

0.78

 raided

0.78

 oversees

0.76

 overseeing

0.75

 issued

0.75

 alleges

0.74

Activations Density 0.050%

government or law-related entities or positions

No Comments

No Known Activations