INDEX

Explanations

adjectives or adverbs that emphasize a positive or negative opinion towards a subject

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

ortmund

-0.79

regate

-0.77

doms

-0.74

idences

-0.73

 Increases

-0.69

 Flavoring

-0.68

icts

-0.66

 greets

-0.66

 scares

-0.66

çīĪ

-0.65

POSITIVE LOGITS

 supposed

1.05

 considered

0.97

 able

0.95

 capable

0.92

 regarded

0.88

 going

0.88

 aware

0.88

 concerned

0.87

 reliant

0.87

 accessible

0.86

Activations Density 1.192%

adjectives or adverbs that emphasize a positive or negative opinion towards a subject

No Comments

No Known Activations