INDEX

Explanations

terms related to political news and government actions

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

Ded

-0.63

natureconservancy

-0.62

Balt

-0.61

 partName

-0.59

nces

-0.58

arij

-0.58

 journal

-0.58

 horm

-0.57

 obser

-0.56

ascript

-0.55

POSITIVE LOGITS

*.

0.81

 hegemony

0.78

'.

0.77

.''.

0.75

ãĢĤ

0.74

'."

0.73

0.71

 amid

0.69

.*

0.68

.'

0.68

Activations Density 0.439%

terms related to political news and government actions

No Comments

No Known Activations