INDEX

Explanations

names of countries and related entities

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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ulner

-0.71

osure

-0.69

ãĥł

-0.67

apter

-0.65

urate

-0.64

CLOSE

-0.63

uracy

-0.63

 DISTR

-0.62

-0.61

 posit

-0.61

POSITIVE LOGITS

 Airlines

1.11

apore

1.07

 Lumpur

0.94

istani

0.90

iang

0.84

ans

0.83

an

0.83

uru

0.78

 Kong

0.77

 Harbor

0.75

Activations Density 0.068%

names of countries and related entities

No Comments

No Known Activations