INDEX

Explanations

phrases related to making requests or suggestions

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

Lenin

-0.62

ROR

-0.60

asta

-0.60

ugar

-0.57

 Femin

-0.56

 Coffin

-0.56

 Shiva

-0.56

Arab

-0.55

orage

-0.55

%:

-0.54

POSITIVE LOGITS

forth

0.95

 understandably

0.92

 temptation

0.91

 naturally

0.89

 logically

0.83

 chances

0.83

 needless

0.83

 shouldn

0.83

 hopefully

0.78

pmwiki

0.76

Activations Density 2.994%

phrases related to making requests or suggestions

No Comments

No Known Activations