INDEX

Explanations

phrases related to trade agreements or political statements

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

theless

-0.82

ĨĴ

-0.81

lihood

-0.71

shire

-0.68

es

-0.67

earch

-0.66

 Emin

-0.63

¬¼

-0.63

erver

-0.62

 Zimmer

-0.62

POSITIVE LOGITS

iffs

1.12

 sands

1.07

zan

0.93

ãĥ£

0.91

balls

0.88

iff

0.87

ball

0.85

thur

0.85

ried

0.84

raz

0.83

Activations Density 0.015%

phrases related to trade agreements or political statements

No Comments

No Known Activations