INDEX

Explanations

details about anime-related events and characters

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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

 Zucker

-0.80

Romney

-0.75

ackets

-0.75

Article

-0.73

 MSNBC

-0.72

Roberts

-0.72

 bumper

-0.71

orters

-0.71

Sports

-0.71

berman

-0.70

POSITIVE LOGITS

 sorce

1.17

 sorcerer

1.14

 treacher

1.10

 corrupted

1.09

 Sorce

1.07

 Norn

1.06

 Celestial

1.03

 Aether

1.03

 cursed

1.02

 mercenaries

1.01

Activations Density 1.347%

details about anime-related events and characters

No Comments

No Known Activations