INDEX

Explanations

names related to storytelling or literature

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

 Lauder

-0.69

 Samar

-0.69

 Lebanese

-0.69

 Dise

-0.66

 Afgh

-0.65

perse

-0.63

WHERE

-0.63

 Zimbabwe

-0.62

 Carbuncle

-0.61

````

-0.60

POSITIVE LOGITS

acious

1.20

itone

1.02

acity

1.01

aic

0.94

acus

0.92

ids

0.90

hett

0.87

ename

0.87

ett

0.87

mund

0.85

Activations Density 0.017%

names related to storytelling or literature

No Comments

No Known Activations