INDEX

Explanations

phrases related to sequential events or orders

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

toe

-0.70

than

-0.64

hab

-0.63

md

-0.62

orum

-0.60

BAT

-0.60

scription

-0.59

aez

-0.59

cos

-0.58

 belief

-0.58

POSITIVE LOGITS

 proceeded

1.22

 proceed

0.97

 promptly

0.84

 abruptly

0.83

 secondly

0.80

 succumb

0.80

 proceeds

0.80

 suddenly

0.76

igslist

0.73

 disappears

0.72

Activations Density 0.968%

phrases related to sequential events or orders

No Comments

No Known Activations