INDEX

Explanations

mentions of high-profile individuals in news or media contexts, particularly related to commentary or media interactions

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

 ILCS

-0.70

chrome

-0.67

 infect

-0.65

 thri

-0.64

 breeds

-0.63

 decay

-0.63

 contamin

-0.63

 disproportion

-0.63

usercontent

-0.62

 depends

-0.62

POSITIVE LOGITS

 keynote

1.30

 conference

1.18

 podium

1.14

 convened

1.08

 attendees

1.04

 unveiling

1.04

 applause

1.03

 televised

1.01

 ceremony

1.01

posium

1.01

Activations Density 4.269%

mentions of high-profile individuals in news or media contexts, particularly related to commentary or media interactions

No Comments

No Known Activations