INDEX

Explanations

conditional statements expressing a likelihood or preference

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

Negative Logits

 Xuan

-0.64

Kag

-0.63

 Chal

-0.60

è¦ļéĨĴ

-0.58

HT

-0.57

 Moving

-0.57

{*

-0.57

 Case

-0.57

Tuc

-0.57

SIG

-0.56

POSITIVE LOGITS

be

1.11

ĸļ

0.94

 suffice

0.92

 prefer

0.92

 gladly

0.92

 ideally

0.91

've

0.88

 argue

0.86

 characterize

0.85

 allow

0.85

Activations Density 1.967%

conditional statements expressing a likelihood or preference

No Known Activations