INDEX

Explanations

proper nouns and technical terms related to specific programs or individuals

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

·

-0.66

absolute

-0.62

amus

-0.62

\\\\\\\\

-0.61

Īè

-0.60

bare

-0.60

gaard

-0.59

ylon

-0.59

 Decay

-0.58

anon

-0.58

POSITIVE LOGITS

ties

1.24

ised

1.21

izations

1.04

isations

1.01

ities

1.00

ty

0.91

ized

0.89

isable

0.88

ity

0.87

ises

0.87

Activations Density 1.736%

proper nouns and technical terms related to specific programs or individuals

No Comments

No Known Activations