INDEX

Explanations

instances of the word "many" with high activation levels

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

UAL

-0.78

anut

-0.77

acus

-0.73

istan

-0.72

ade

-0.70

oulos

-0.68

atum

-0.67

anything

-0.66

etry

-0.66

icism

-0.65

POSITIVE LOGITS

 times

1.10

 facets

1.03

body

0.96

 aspects

0.94

 unanswered

0.87

 thousands

0.85

ones

0.83

 misconceptions

0.82

 similarities

0.82

 instances

0.82

Activations Density 0.407%

instances of the word "many" with high activation levels

No Comments

No Known Activations