INDEX

Explanations

descriptive phrases related to a negative assessment or criticism

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

catentry

-0.89

Recommend

-0.79

ItemImage

-0.77

ourses

-0.76

ructure

-0.75

Import

-0.75

ilitation

-0.74

isSpecialOrderable

-0.73

ODUCT

-0.72

Initially

-0.69

POSITIVE LOGITS

 gigg

1.16

 dudes

1.07

 jokes

1.06

 boobs

1.05

 nerds

1.05

 dick

1.05

 shit

1.03

 fries

1.02

 booze

1.02

 crap

1.01

Activations Density 6.886%

descriptive phrases related to a negative assessment or criticism

No Comments

No Known Activations