INDEX

Explanations

political and policy-related terms and actions

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 partName

-0.81

mist

-0.68

spot

-0.68

Eva

-0.67

Sym

-0.66

 Frie

-0.65

 sorrow

-0.65

 Correspond

-0.65

 ERROR

-0.65

 moments

-0.64

POSITIVE LOGITS

 decriminal

1.16

 privat

1.16

 abol

1.14

 legalized

1.13

 privatization

1.11

 decentral

1.10

 reforms

1.09

 abolish

1.06

 abolition

1.03

 overhaul

1.03

Activations Density 4.916%

political and policy-related terms and actions

No Comments

No Known Activations