INDEX

Explanations

expressions of personal opinions or desires in various contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Xuan

-0.64

Kag

-0.63

è¦ļéĨĴ

-0.62

 Chal

-0.61

 Trap

-0.58

HT

-0.57

 Moving

-0.57

 Case

-0.56

Vis

-0.56

SIG

-0.56

POSITIVE LOGITS

be

1.13

 gladly

0.96

 prefer

0.95

ĸļ

0.92

 suffice

0.91

 dearly

0.90

 ideally

0.90

 characterize

0.87

 doubtless

0.86

've

0.86

Activations Density 0.725%

expressions of personal opinions or desires in various contexts

No Comments

No Known Activations