INDEX

Explanations

proper nouns and names of companies or organizations

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

 aisle

-0.74

 stalk

-0.73

 carrier

-0.69

 carriers

-0.68

 rush

-0.65

 trademark

-0.62

++++++++++++++++

-0.62

 wheelchair

-0.62

 generals

-0.62

 theorist

-0.61

POSITIVE LOGITS

acia

0.79

nergy

0.69

Keys

0.69

icro

0.68

mad

0.68

ild

0.68

atum

0.67

nil

0.67

WARN

0.67

Nap

0.67

Activations Density 0.223%

proper nouns and names of companies or organizations

No Comments

No Known Activations