INDEX

Explanations

instructions or steps in a giveaway or contest

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

arious

-0.76

][

-0.67

ensive

-0.65

 Originally

-0.64

istar

-0.64

Iraq

-0.64

uper

-0.63

mosp

-0.63

agen

-0.62

existence

-0.62

POSITIVE LOGITS

 preferably

1.17

 then

1.15

 enjoy

1.12

 THEN

1.07

 proceed

1.06

 reap

1.05

 beware

1.04

 apply

1.03

 prest

1.02

vo

1.02

Activations Density 0.166%

instructions or steps in a giveaway or contest

No Comments

No Known Activations