INDEX

Explanations

mention of specific government and healthcare terms

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

ãĥł

-0.73

ãĤ·ãĥ£

-0.68

antha

-0.68

gered

-0.66

icio

-0.65

rar

-0.65

CVE

-0.65

ger

-0.64

catentry

-0.64

iesta

-0.64

POSITIVE LOGITS

ION

0.93

IONS

0.91

MEN

0.89

ING

0.87

 MARK

0.85

ONDON

0.83

VILLE

0.83

IAN

0.82

WORK

0.82

TO

0.81

Activations Density 0.013%

mention of specific government and healthcare terms

No Comments

No Known Activations