INDEX

Explanations

entities or terms related to political or legal matters

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 polyg

-0.72

 scatter

-0.67

 sacrific

-0.67

 beaut

-0.66

 decomp

-0.64

nod

-0.63

 charm

-0.62

RAD

-0.62

 scattering

-0.61

 maximum

-0.61

POSITIVE LOGITS

¬

1.07

¹

1.05

Â¢

1.03

ı

1.03

¡

1.02

£

1.00

Ĵ

0.97

¤

0.92

²

0.92

¸

0.91

Activations Density 1.385%

entities or terms related to political or legal matters

No Comments

No Known Activations