INDEX

Explanations

modal verbs indicating suggestion or obligation

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

Fra

-0.79

Wid

-0.68

WI

-0.65

 Puzzle

-0.65

 Hilbert

-0.62

 Syndrome

-0.61

ROS

-0.59

atile

-0.59

 Afgh

-0.59

Kah

-0.58

POSITIVE LOGITS

ered

1.15

be

1.06

nt

1.01

 ideally

0.96

0.94

ering

0.94

 strive

0.90

 aspire

0.90

 beware

0.87

ublic

0.87

Activations Density 1.112%

modal verbs indicating suggestion or obligation

No Comments

No Known Activations