INDEX

Explanations

proper nouns/terms related to legal or governmental entities

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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ashtra

-0.87

 dogs

-0.73

 goats

-0.73

 verses

-0.71

 jars

-0.69

abis

-0.68

ibling

-0.68

aways

-0.66

 springs

-0.66

respective

-0.65

POSITIVE LOGITS

 Personnel

1.03

 Strategic

1.02

 Communications

0.94

 Inspector

0.94

 Management

0.92

 Budget

0.89

 Inquiry

0.89

CLASSIFIED

0.87

 Naval

0.83

 Information

0.81

Activations Density 0.075%

proper nouns/terms related to legal or governmental entities

No Comments

No Known Activations

proper nouns/terms related to legal or governmental entities

No Comments

No Known Activations