INDEX

Explanations

phrases related to events or achievements

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

 Sleeping

-0.61

 Penguin

-0.59

 Tayyip

-0.59

 punitive

-0.58

 Scholar

-0.57

 Preferences

-0.56

ROM

-0.56

 deployments

-0.56

 Household

-0.56

 combustion

-0.55

POSITIVE LOGITS

ividual

1.47

iversity

1.12

rew

1.12

yssey

1.01

hyde

0.98

onna

0.98

ynamic

0.97

iscover

0.97

iamond

0.96

etermined

0.95

Activations Density 0.038%

phrases related to events or achievements

No Comments

No Known Activations