INDEX

Explanations

ordinal numbers indicating position in a sequence

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

Statistics

-0.70

mbuds

-0.69

bos

-0.68

uploads

-0.63

 Monsters

-0.61

Downloadha

-0.61

bda

-0.60

info

-0.60

Attributes

-0.59

values

-0.59

POSITIVE LOGITS

 inning

1.05

ousand

1.02

irty

1.01

 century

1.00

 installment

0.97

 Amendment

0.97

 baseman

0.96

 percentile

0.93

century

0.89

 harmonic

0.89

Activations Density 0.709%

ordinal numbers indicating position in a sequence

No Comments

No Known Activations