INDEX

Explanations

phrases related to consequences or impacts

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

ahime

-0.75

eg

-0.73

ortment

-0.71

rongh

-0.69

ometimes

-0.67

hap

-0.61

itton

-0.60

eele

-0.59

initely

-0.58

eworks

-0.57

POSITIVE LOGITS

 whatsoever

2.20

nor

1.48

 anymore

1.20

 except

1.07

nor

1.04

 slightest

0.96

soever

0.91

 anywhere

0.88

 anybody

0.86

 anything

0.85

Activations Density 1.642%

phrases related to consequences or impacts

No Comments

No Known Activations