INDEX

Explanations

legal terms and court-related phrases

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

jiang

-0.72

ku

-0.72

mson

-0.68

 Saban

-0.67

 Pyro

-0.66

ï¸ı

-0.65

resso

-0.63

Ã´

-0.63

 Horus

-0.61

tz

-0.60

POSITIVE LOGITS

rooms

1.26

esan

1.04

room

1.03

 proceedings

0.95

auld

0.95

 rulings

0.93

 justices

0.93

 hearings

0.91

 judge

0.86

Court

0.86

Activations Density 0.341%

legal terms and court-related phrases

No Comments

No Known Activations