INDEX

Explanations

phrases that involve an emotional or intentional aspect

oai_token-act-pair · gpt-3.5-turbo

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

 ÂŃ

-0.68

à©

-0.66

 iCloud

-0.61

 à¨

-0.61

SAP

-0.61

 FedEx

-0.60

 crypt

-0.60

 ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½ï¿½

-0.60

 autop

-0.58

 Guant

-0.58

POSITIVE LOGITS

pedia

0.91

utenberg

0.81

friends

0.76

him

0.75

ulo

0.73

values

0.71

iland

0.71

lyn

0.70

ilts

0.67

eda

0.67

Activations Density 0.732%

phrases that involve an emotional or intentional aspect

No Comments

No Known Activations