INDEX

Explanations

phrases related to instructions and commands

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

osa

-0.70

osion

-0.64

 Crus

-0.63

 Tanz

-0.63

ologically

-0.61

å£«

-0.61

manager

-0.59

alyses

-0.58

Subject

-0.58

itu

-0.58

POSITIVE LOGITS

theless

1.19

 been

1.14

been

1.09

 Been

1.05

gotten

0.88

 gotten

0.83

 gotta

0.82

mber

0.81

got

0.80

tta

0.80

Activations Density 0.693%

phrases related to instructions and commands

No Comments

No Known Activations