INDEX

Explanations

phrases or references related to the idea of displaying or presenting information

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Dame

-0.64

 hurd

-0.60

uty

-0.58

 litter

-0.57

 tatt

-0.56

eco

-0.56

 torches

-0.56

 mascul

-0.56

 contempl

-0.56

 psycho

-0.56

POSITIVE LOGITS

biz

0.99

 Thumbnails

0.92

case

0.91

downs

0.79

cases

0.79

anooga

0.78

iao

0.72

hide

0.70

Alert

0.69

Reviewer

0.67

Activations Density 0.005%

phrases or references related to the idea of displaying or presenting information

No Comments

No Known Activations