INDEX

Explanations

mentions of software features or updates

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 recalls

-0.62

ecd

-0.59

etting

-0.58

etter

-0.58

 eyebrow

-0.56

 Quote

-0.56

cffffcc

-0.56

 believer

-0.55

 caveat

-0.55

Interested

-0.54

POSITIVE LOGITS

 been

1.35

 gone

1.18

 gotten

1.16

been

1.15

 become

1.09

 fallen

1.05

 risen

1.05

 undergone

1.03

 begun

1.02

 Been

1.01

Activations Density 0.299%

mentions of software features or updates

No Comments

No Known Activations