INDEX

Explanations

phrases related to politics and public opinion

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

aten

-0.76

atcher

-0.71

umblr

-0.70

 Rebell

-0.63

 Cycle

-0.61

 Olympus

-0.57

edIn

-0.57

 discarded

-0.57

 Seas

-0.57

 PAGE

-0.56

POSITIVE LOGITS

 insight

1.02

 insights

0.84

 flexibility

0.83

 assurance

0.83

 valuable

0.83

 benefits

0.80

 useful

0.79

icum

0.78

 advantages

0.78

 expertise

0.75

Activations Density 1.669%

phrases related to politics and public opinion

No Comments

No Known Activations