INDEX

Explanations

phrases related to political commentary or opinions

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

adra

-0.74

catentry

-0.69

Availability

-0.68

CTR

-0.65

ante

-0.64

 availability

-0.61

Campaign

-0.60

reditation

-0.59

BuyableInstoreAndOnline

-0.57

Requ

-0.57

POSITIVE LOGITS

 aloud

1.10

uyomi

0.83

 provoking

0.73

evaluate

0.72

swer

0.71

borgh

0.70

gaard

0.69

 hypot

0.66

 nostalg

0.66

 contemplating

0.65

Activations Density 0.117%

phrases related to political commentary or opinions

No Comments

No Known Activations