INDEX

Explanations

concepts or phrases related to contrasting opinions or stances on an issue

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_fs6144-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_6144

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

6,144

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

nce

-0.69

pta

-0.61

rupted

-0.60

utory

-0.59

 Newsletter

-0.59

 Rapids

-0.58

ķ

-0.58

ogle

-0.58

Und

-0.57

kie

-0.57

POSITIVE LOGITS

 sexes

1.25

 alike

1.20

 equally

1.19

 simultaneously

1.15

 sides

1.05

 mutually

1.02

 concurrently

0.97

 halves

0.97

 vying

0.95

 genders

0.92

Activations Density 0.580%

concepts or phrases related to contrasting opinions or stances on an issue

No Comments

No Known Activations