INDEX

Explanations

locations or street names

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

hower

-0.74

hound

-0.73

 "$:/

-0.70

ļéĨĴ

-0.70

EStream

-0.62

 friendly

-0.60

 thumbs

-0.60

 limited

-0.59

 deaf

-0.58

ãĥ¼ãĥ³

-0.58

POSITIVE LOGITS

rict

1.22

alker

1.18

amped

1.14

uffed

1.12

onew

1.10

itched

1.10

okes

1.10

oppable

1.06

amps

1.06

amping

1.06

Activations Density 0.422%

locations or street names

No Comments

No Known Activations