INDEX

Explanations

personal pronouns and words related to personal reflections or internal struggles

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

 Eleven

-0.63

 Outside

-0.60

 Joint

-0.58

Jub

-0.58

 Comic

-0.56

 Dunk

-0.56

Outside

-0.56

marked

-0.55

 Middles

-0.55

 Ranking

-0.54

POSITIVE LOGITS

'll

1.61

're

1.35

've

1.15

can

1.13

'd

1.12

 shouldn

1.02

 might

0.99

 will

0.99

 could

0.99

don

0.94

Activations Density 0.367%

personal pronouns and words related to personal reflections or internal struggles

No Comments

No Known Activations