INDEX

Explanations

phrases related to expressing opinions or facts

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

 crunch

-0.59

 lungs

-0.57

ALS

-0.57

$$

-0.57

 butterfly

-0.55

ses

-0.55

eded

-0.55

mul

-0.54

GBT

-0.54

OK

-0.53

POSITIVE LOGITS

ername

0.83

entimes

0.80

renheit

0.77

olkien

0.76

ournals

0.71

oln

0.71

rophe

0.69

anship

0.67

enza

0.66

terday

0.65

Activations Density 0.447%

phrases related to expressing opinions or facts

No Comments

No Known Activations