INDEX

Explanations

terms related to linguistics and language studies

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

ãĥ£

-0.68

osaurs

-0.65

minecraft

-0.62

Minecraft

-0.62

Flickr

-0.61

ertain

-0.59

éĸ

-0.58

cam

-0.58

ãĤµ

-0.57

natureconservancy

-0.57

POSITIVE LOGITS

rul

0.71

rique

0.69

eers

0.68

 equivalents

0.64

 withd

0.61

 skelet

0.58

 Azerb

0.56

 minus

0.56

LM

0.55

alion

0.55

Activations Density 6.135%

terms related to linguistics and language studies

No Comments

No Known Activations