INDEX

Explanations

phrases related to discussions or statements about specific topics

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

icons

-0.84

pps

-0.78

asters

-0.77

ãĥĥ

-0.76

orians

-0.76

ormons

-0.73

å§«

-0.72

 Roads

-0.70

asts

-0.69

arbon

-0.69

POSITIVE LOGITS

 pesky

1.20

 particular

1.11

 kind

1.03

 elusive

1.02

 sort

0.98

 fateful

0.96

 sucker

0.94

 type

0.93

 aspect

0.92

 equation

0.90

Activations Density 1.382%

phrases related to discussions or statements about specific topics

No Comments

No Known Activations