INDEX

Explanations

concepts related to philosophy, religion, and political systems

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

":"/

-0.78

._

-0.62

ady

-0.62

imester

-0.62

 guiName

-0.61

worldly

-0.60

owe

-0.58

İ

-0.57

":-

-0.56

uder

-0.56

POSITIVE LOGITS

etc

1.99

etc

1.69

ect

1.06

whatever

0.92

et

0.88

and

0.82

...)

0.78

âĢ¦)

0.75

ĪĴ

0.70

Activations Density 0.807%

concepts related to philosophy, religion, and political systems

No Comments

No Known Activations