INDEX

Explanations

references to government departments or legal proceedings

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

jad

-0.68

 arteries

-0.68

Ô

-0.66

 scenarios

-0.65

 caric

-0.64

 traces

-0.63

NetMessage

-0.63

 outnumbered

-0.62

 verses

-0.61

chall

-0.60

POSITIVE LOGITS

 Agriculture

1.17

 Commerce

1.15

 Transportation

1.09

 Interior

0.99

 Justice

0.97

 Correction

0.95

 Corrections

0.93

 Revenue

0.93

 Homeland

0.93

 Communications

0.90

Activations Density 0.485%

references to government departments or legal proceedings

No Comments

No Known Activations