INDEX

Explanations

proper nouns related to TV shows, events, and personalities

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

aff

-0.77

bon

-0.77

ordering

-0.74

uga

-0.74

medium

-0.73

bring

-0.73

terms

-0.72

agara

-0.71

Äĩ

-0.70

priv

-0.70

POSITIVE LOGITS

 Comics

1.01

 Shooter

0.98

 Theory

0.97

 Arcade

0.96

 Ninja

0.96

 Productions

0.96

 Blind

0.94

 Brewing

0.93

 Initiative

0.93

 Trading

0.93

Activations Density 0.160%

proper nouns related to TV shows, events, and personalities

No Comments

No Known Activations