INDEX

Explanations

references to specific publications like "The New York Times" and "Los Angeles Times" in different contexts

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

sed

-0.83

cles

-0.74

gone

-0.67

unts

-0.66

venge

-0.64

ardless

-0.63

bands

-0.63

shift

-0.63

isable

-0.63

together

-0.62

POSITIVE LOGITS

 Editorial

1.27

 editorial

1.23

 Newsp

1.20

 Magazine

1.18

 columnist

1.15

 reporter

1.10

 Newspaper

1.09

 Editors

1.09

 reporters

1.06

 editors

1.05

Activations Density 1.560%

references to specific publications like "The New York Times" and "Los Angeles Times" in different contexts

No Comments

No Known Activations