INDEX

Explanations

phrases related to legal and political matters, including discussions about actions, laws, rights, investigations, and justice outcomes

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

 prelim

-0.65

 conceal

-0.62

 assassin

-0.61

 retri

-0.60

 killer

-0.60

pit

-0.59

 moth

-0.59

ombo

-0.59

 coron

-0.58

 fractions

-0.58

POSITIVE LOGITS

 Specifically

1.20

Specifically

1.14

Refer

1.12

 Asked

1.11

 Speaking

1.07

He

1.05

``

1.05

 Saying

1.05

Speaking

1.03

Calling

1.02

Activations Density 1.845%

phrases related to legal and political matters, including discussions about actions, laws, rights, investigations, and justice outcomes

No Comments

No Known Activations