INDEX

Explanations

words related to crisis or emergencies

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

alty

-0.92

alties

-0.81

rosse

-0.80

ipedia

-0.78

icrobial

-0.78

ells

-0.78

sburg

-0.76

Īè

-0.76

eele

-0.75

rica

-0.74

POSITIVE LOGITS

 aver

1.04

 plag

1.02

 engulf

1.01

 crisis

0.97

 unfolding

0.88

 epidemic

0.87

 crises

0.87

 involving

0.83

 gripping

0.83

 arising

0.81

Activations Density 0.025%

words related to crisis or emergencies

No Comments

No Known Activations