INDEX

Explanations

software-related terms and acronyms, possibly related to code or development

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

fulness

-0.83

baugh

-0.71

coni

-0.69

inating

-0.64

 Nadu

-0.61

AFP

-0.61

nings

-0.60

inately

-0.58

 concentration

-0.58

ographically

-0.58

POSITIVE LOGITS

atari

1.03

ois

0.99

naire

0.98

WER

0.96

ues

0.94

atar

0.94

addafi

0.93

UE

0.93

idian

0.89

eros

0.87

Activations Density 1.849%

software-related terms and acronyms, possibly related to code or development

No Comments

No Known Activations