INDEX

Explanations

questions ending with a question mark

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

 background

-0.69

 marsh

-0.67

ality

-0.66

dex

-0.63

aper

-0.63

arm

-0.63

 swamp

-0.61

 corrid

-0.61

aku

-0.61

 evening

-0.61

POSITIVE LOGITS

 Nope

1.34

 Surely

1.27

 Answer

1.26

 Well

1.26

 Probably

1.22

 Certainly

1.18

 Wouldn

1.16

Why

1.15

ï¿½ï¿½ï¿½ï¿½

1.14

Answer

1.12

Activations Density 0.409%

questions ending with a question mark

No Comments

No Known Activations