INDEX

Explanations

predictions or potential outcomes

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

BN

-0.70

tained

-0.69

Detailed

-0.68

leground

-0.67

OWN

-0.66

FP

-0.65

pec

-0.64

ivered

-0.63

oph

-0.63

nai

-0.62

POSITIVE LOGITS

 chances

1.10

then

1.06

 then

1.02

 THEN

0.91

it

0.87

 they

0.83

 surely

0.82

 congratulations

0.81

 please

0.80

 however

0.76

Activations Density 0.907%

predictions or potential outcomes

No Comments

No Known Activations