INDEX

Explanations

phrases related to governmental organizations and official reports

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

 classmates

-0.76

 strips

-0.73

 dots

-0.73

fro

-0.73

 pods

-0.71

brainer

-0.71

 scrim

-0.69

 classmate

-0.68

 flashbacks

-0.66

aciously

-0.66

POSITIVE LOGITS

 Agriculture

1.24

 Computing

1.17

 Management

1.17

 Arts

1.12

 Communication

1.11

 Communications

1.11

 Medicine

1.10

 Technology

1.08

 Aging

1.07

 Personnel

1.06

Activations Density 3.626%

phrases related to governmental organizations and official reports

No Comments

No Known Activations