INDEX

Explanations

proper nouns and terms related to legal issues and political events

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

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

drive

-0.71

body

-0.68

fax

-0.68

eur

-0.67

warming

-0.63

oler

-0.61

rip

-0.61

ebook

-0.60

ruption

-0.60

ebus

-0.60

POSITIVE LOGITS

 grouped

1.21

 listed

1.07

 include

1.00

 numbered

0.97

 collectively

0.96

:[

0.94

 namely

0.93

 clustered

0.93

listed

0.92

:-

0.92

Activations Density 0.541%

proper nouns and terms related to legal issues and political events

No Comments

No Known Activations