INDEX

Explanations

phrases related to expediting or accelerating processes, especially administrative or regulatory ones

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

ukong

-0.85

gdala

-0.83

Champ

-0.72

loo

-0.71

entric

-0.68

owship

-0.68

kok

-0.68

uably

-0.66

ãĤ¤ãĥĪ

-0.66

chens

-0.65

POSITIVE LOGITS

 pace

1.24

 paced

1.10

 turnaround

0.97

 progress

0.96

 rollout

0.95

 onset

0.94

 processing

0.94

 digestion

0.89

 commute

0.89

 transition

0.88

Activations Density 0.156%

phrases related to expediting or accelerating processes, especially administrative or regulatory ones

No Comments

No Known Activations