INDEX

Explanations

people's names or terms related to politics and news channels, especially CNN and Fox News

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

ovych

-0.70

tip

-0.64

schild

-0.63

kil

-0.61

 impulse

-0.61

Â¯

-0.59

WARD

-0.58

gat

-0.58

idays

-0.57

yip

-0.57

POSITIVE LOGITS

ulhu

1.16

estial

0.88

berus

0.84

ãĤ¼ãĤ¦ãĤ¹

0.83

ioxide

0.82

igham

0.80

ornia

0.78

emporary

0.77

estine

0.74

pillar

0.72

Activations Density 1.323%

people's names or terms related to politics and news channels, especially CNN and Fox News

No Comments

No Known Activations