INDEX

Explanations

references to proper nouns and technical terms related to scientific models and geographic locations

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

æĸ¹

-0.83

 Debor

-0.78

 Radius

-0.65

ģ

-0.65

ãĥ¼ãĥĨãĤ£

-0.64

ItemTracker

-0.61

 Picture

-0.60

ellen

-0.58

eers

-0.58

ELS

-0.57

POSITIVE LOGITS

anguage

1.17

ibrary

1.16

oaded

1.16

uxe

1.07

ength

1.05

ounge

1.01

anguages

1.00

utive

1.00

uvian

1.00

iberal

1.00

Activations Density 0.773%

references to proper nouns and technical terms related to scientific models and geographic locations

No Comments

No Known Activations