INDEX

Explanations

phrases related to skepticism or criticism

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

catentry

-0.69

 assorted

-0.65

ials

-0.59

 Locations

-0.59

 pitch

-0.58

idelines

-0.57

 periodically

-0.56

bil

-0.55

 Reported

-0.54

ortment

-0.53

POSITIVE LOGITS

 anymore

1.13

nor

1.00

 anybody

0.86

 EVER

0.83

 whatsoever

0.80

 unless

0.74

 anyone

0.73

 anything

0.73

 bothered

0.72

unless

0.72

Activations Density 1.863%

phrases related to skepticism or criticism

No Comments

No Known Activations