INDEX

Explanations

explicit and strong language

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_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

Dex

-0.73

Recomm

-0.70

Decre

-0.70

Dialog

-0.69

PsyNetMessage

-0.67

Broad

-0.66

 obser

-0.66

BIL

-0.65

Ctrl

-0.65

 easing

-0.64

POSITIVE LOGITS

holes

1.26

fuck

1.20

hole

1.15

ishly

1.02

shit

0.99

dump

0.93

loads

0.91

amn

0.90

nuts

0.90

wit

0.84

Activations Density 1.221%

explicit and strong language

No Comments

No Known Activations