INDEX

Explanations

names and titles of people, especially in a professional setting

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

interrupted

-0.74

antha

-0.67

etheless

-0.63

 sweep

-0.62

 bottleneck

-0.62

ailability

-0.62

ipolar

-0.61

 pressures

-0.61

rha

-0.61

olicy

-0.60

POSITIVE LOGITS

brate

1.54

brates

1.50

ller

1.21

llers

1.19

levision

1.18

achers

1.11

llo

1.09

lla

1.07

achable

1.05

lli

1.05

Activations Density 0.023%

names and titles of people, especially in a professional setting

No Comments

No Known Activations