INDEX

Explanations

references to political events or governmental actions and 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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

icons

-0.91

acers

-0.76

Examples

-0.73

masters

-0.73

okers

-0.72

lee

-0.72

jobs

-0.70

mates

-0.70

loads

-0.70

master

-0.68

POSITIVE LOGITS

 century

1.05

 hemisphere

1.01

 particular

1.01

decade

0.95

 year

0.95

 venerable

0.95

 country

0.94

 week

0.92

 regard

0.92

 endeavor

0.90

Activations Density 0.188%

references to political events or governmental actions and statements

No Comments

No Known Activations