INDEX

Explanations

proper nouns and titles

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

ounter

-0.75

icably

-0.74

 allowable

-0.73

ities

-0.69

etheless

-0.69

 toler

-0.68

 proble

-0.67

ITIES

-0.67

ictions

-0.67

ancies

-0.67

POSITIVE LOGITS

love

1.02

Works

1.01

 Squad

0.98

Maker

0.98

 Breaker

0.97

 Runner

0.96

Squ

0.96

breaker

0.96

Point

0.95

 Girl

0.94

Activations Density 1.713%

proper nouns and titles

No Comments

No Known Activations