INDEX

Explanations

terms related to healthcare, private services, and health policies

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

iple

-0.83

urat

-0.79

enance

-0.72

inal

-0.72

anism

-0.71

ength

-0.70

ilts

-0.70

 Roof

-0.68

naire

-0.67

ima

-0.67

POSITIVE LOGITS

 agency

1.20

 agencies

1.20

policy

1.17

 department

1.16

 ministry

1.15

 bureau

1.13

 advisory

1.11

 policy

1.07

 ministries

1.03

 consultancy

1.03

Activations Density 4.280%

terms related to healthcare, private services, and health policies

No Comments

No Known Activations