INDEX

Explanations

names in news articles, particularly focusing on legal or controversial 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_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

 EDITION

-0.72

 Thumbnails

-0.69

AAA

-0.64

Reloaded

-0.63

CLS

-0.62

 Destination

-0.60

RIS

-0.60

*.

-0.59

 Spaces

-0.59

PBS

-0.59

POSITIVE LOGITS

enson

1.07

leck

0.98

auer

0.97

acci

0.96

aney

0.95

jan

0.94

kowski

0.94

inson

0.94

linger

0.91

inski

0.91

Activations Density 1.143%

names in news articles, particularly focusing on legal or controversial contexts

No Comments

No Known Activations