INDEX

Explanations

the repeated word "ha."

oai_token-act-pair · gpt-4o-mini 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

atories

-0.76

rations

-0.69

 outgoing

-0.68

bats

-0.68

rats

-0.63

entric

-0.63

inarily

-0.62

 Collider

-0.61

 Reson

-0.61

é»Ĵ

-0.61

POSITIVE LOGITS

wn

1.27

user

1.17

illard

0.99

irst

0.93

pless

0.93

pta

0.92

ppa

0.92

uthor

0.91

fter

0.91

emi

0.90

Activations Density 0.016%

the repeated word "ha."

No Comments

No Known Activations