INDEX

Explanations

mentions of strategies and evaluations in political and social contexts

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

sonian

-0.78

interrupted

-0.67

spin

-0.66

 Seen

-0.66

bek

-0.65

lite

-0.65

Champ

-0.64

 shalt

-0.64

maj

-0.64

Gi

-0.64

POSITIVE LOGITS

 regards

1.49

 regard

1.36

 handling

1.15

 terms

1.11

 tackling

0.99

 shaping

0.98

 crafting

0.98

 selecting

0.98

 determining

0.97

 navigating

0.97

Activations Density 3.636%

mentions of strategies and evaluations in political and social contexts

No Comments

No Known Activations