INDEX

Explanations

phrases containing email addresses

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Opportun

-1.00

 Accountability

-1.00

 Parish

-0.96

 Regulations

-0.96

 Ninth

-0.95

 Guarant

-0.93

 Lines

-0.92

 Excellence

-0.92

 Barrier

-0.91

 Parties

-0.90

POSITIVE LOGITS

1.20

info

1.09

hd

1.08

cli

1.05

xp

1.05

dc

1.04

wolf

1.03

pd

1.02

rx

1.02

dragon

1.01

Activations Density 2.634%

phrases containing email addresses

No Comments

No Known Activations