INDEX

Explanations

phrases related to negation or uncertainty

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

ibaba

-0.96

ovsky

-0.67

 Pengu

-0.65

abi

-0.62

ashtra

-0.62

vana

-0.61

etsk

-0.61

ario

-0.60

swers

-0.58

oided

-0.57

POSITIVE LOGITS

 margins

1.06

 margin

1.06

 means

0.95

 virtue

0.87

 leaps

0.85

 standards

0.82

 Means

0.81

 degrees

0.76

committee

0.71

dB

0.70

Activations Density 0.149%

phrases related to negation or uncertainty

No Comments

No Known Activations