INDEX

Explanations

words related to political and legislative activities, including terms related to government processes and officials

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

RTX

-0.63

 Families

-0.63

 Defendants

-0.62

 Conversation

-0.62

 Shang

-0.60

ancial

-0.58

 Nurs

-0.57

 Definitions

-0.56

 mortality

-0.55

CPU

-0.55

POSITIVE LOGITS

è£ħ

0.77

 insisted

0.74

 esche

0.73

.,

0.70

 nurt

0.70

 opted

0.70

 grew

0.69

 vehemently

0.69

 contends

0.68

 turned

0.68

Activations Density 0.052%

words related to political and legislative activities, including terms related to government processes and officials

No Comments

No Known Activations