INDEX

Explanations

words related to governance, politics, and societal issues

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

chwitz

-0.69

andise

-0.68

 Indra

-0.67

ADRA

-0.66

reon

-0.65

CLASSIFIED

-0.63

HEAD

-0.61

ONSORED

-0.59

 indo

-0.58

 hotly

-0.58

POSITIVE LOGITS

(<

1.12

pox

0.99

 minded

0.94

azaki

0.83

minded

0.80

folk

0.80

case

0.80

est

0.80

tiny

0.76

entary

0.76

Activations Density 2.234%

words related to governance, politics, and societal issues

No Comments

No Known Activations