INDEX

Explanations

descriptions of traffic incidents and related comments from people involved

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

 treason

-0.86

 FISA

-0.72

 undermining

-0.72

ius

-0.70

eus

-0.67

Conclusion

-0.67

).[

-0.67

canon

-0.67

 assassination

-0.66

 imperialist

-0.66

POSITIVE LOGITS

 remembers

0.91

 recalls

0.84

 volunteering

0.81

 firsthand

0.80

ivan

0.78

 biking

0.78

 granddaughter

0.77

veland

0.77

 volunteers

0.77

atown

0.77

Activations Density 3.050%

descriptions of traffic incidents and related comments from people involved

No Comments

No Known Activations